Supported Model Frameworks/Formats - KServe
支持的模型框架/格式
- 模型服务运行时间
- 使用推理服务部署Tensorflow模型
-
- 创建HTTP推理服务
-
- 运行预测
- 金丝雀Rollout
- 创建gRPC推理服务
-
- 运行预测
- 使用TorchServe推理服务部署PyTorch模型
-
- 使用模型档案文件和配置创建模型存储
- 使用V1 REST协议部署PyTorch模型
-
- 创建TorchServe推理服务
- 模型推理
- 模型说明
- 使用V2 REST协议部署PyTorch模型
-
- 创建推理服务
- 模型推理
- 模型说明
- 自动缩放
-
- Knative Autoscaler
- 创建具有并发目标的推理服务
- 使用并发请求运行推理
- Check Pod Autoscaling
- 金丝雀发布
-
- 使用Canary模型创建推理服务
- 检查流量状况
- “流量”推出
- 回滚模型
- 监控
- 使用推理服务部署Scikit学习模型
-
- 训练
- 本地测试
-
- 先决条件
- 模型设置
- 本地服务模型
- 使用推理服务进行部署
- 测试部署的模型
- 使用推理服务部署XGBoost模型
-
- 训练
- 本地测试
-
- 先决条件
- 模型设置
- 在本地为我们的模型服务
- 使用推理服务进行部署
- 测试部署的模型
- 使用推理服务部署PMML模型
-
- 创建推理服务
- 运行预测
- 使用PMML推理服务部署Spark MLlib模型
-
- 设置
- 训练Spark MLlib模型并导出到PMML文件
- 使用PMMLServer创建推理服务
-
- 运行预测
- 使用推理服务部署LightGBM模型
-
- 训练LightGBM模型
- 使用V1协议部署LightGBM模型
-
- 在本地测试模型
- 使用推理服务进行部署
- 测试已部署的模型
- 使用[V2协议](https://github.com/kserve/kserve/tree/master/docs/predict-api/v2)部署模型
-
- 在本地测试模型
- 使用REST端点部署推理服务
- 用curl测试展开的模型
- 使用gRPC端点创建推理服务
- 使用grpcurl测试已部署的模型
- 使用推理服务部署Paddle模型
-
- 创建推理服务
- 运行预测
- 使用推理服务部署MLflow模型
-
- 训练
- 本地测试
-
- 先决条件
- 模型设置
- 在本地启动模型服务器
- 使用推理服务进行部署
- 测试部署的模型
https://kserve.github.io/website/0.10/modelserving/v1beta1/serving_runtime/
模型服务运行时间
KServe提供了一个简单的Kubernetes CRD，可以将单个或多个经过训练的模型部署到模型服务运行时，如TFServing、TorchServe、Triton Inference Server。此外，ModelServer是在KServe中使用预测v1协议实现的Python模型服务运行时，MLServer使用REST和gRPC实现预测v2协议。这些模型服务运行时能够提供开箱即用的模型服务，但您也可以选择为更复杂的用例构建自己的模型服务器。KServe提供了基本的API基元，使您可以轻松地构建自定义模型服务运行时，您可以使用BentML等其他工具来构建您的自定义模型服务镜像。
使用推理服务部署模型后，您将获得KServe提供的以下所有无服务器功能。
- 缩放到零和从零缩放
- 基于请求的CPU/GPU自动缩放
- 修订管理
- 优化的容器
- 批处理
- 请求/响应日志记录
- 流量管理
- AuthN/AuthZ的安全性
- 分布式跟踪
- 开箱即用的指标
- 入口/出口控制
下表列出了KServe支持的每个模型服务运行时。HTTP和gRPC列指示服务运行时支持的预测协议版本。KServe预测协议被记为“v1”或“v2”。一些服务运行时也支持它们自己的预测协议，这些协议用*表示。默认的服务运行时版本列定义了服务运行时的源和版本——MLServer、KServe或它自己的。这些版本也可以在runtime kustomization YAML中找到。为运行时提供服务的所有KServe本机模型都使用当前的KServe发布版本（v0.10）。支持的框架版本列列出了支持的模型的主要版本。这些也可以在相应的runtime YAML中supportedModelFormats字段下找到。对于使用KServe服务运行时的模型框架，可以在KServe/python中找到特定的默认版本。在给定的服务运行时目录中，setup.py文件包含所使用的确切模型框架版本。例如，在kserve/python/lgbserver中，setup.py文件将模型框架版本设置为3.3.2，lightgbm==3.3.2。
| Model Serving Runtime | Exported model | HTTP | gRPC | Default Serving Runtime Version | Supported Framework (Major) Version(s) | Examples |
|---|---|---|---|---|---|---|
| Custom ModelServer | – | v1, v2 | v2 | – | – | Custom Model |
| LightGBM MLServer | Saved LightGBM Model | v2 | v2 | v1.0.0 (MLServer) | 3 | LightGBM Iris V2 |
| LightGBM ModelServer | Saved LightGBM Model | v1 | – | v0.10 (KServe) | 3 | LightGBM Iris |
| MLFlow ModelServer | Saved MLFlow Model | v2 | v2 | v1.0.0 (MLServer) | 1 | MLFLow wine-classifier |
| PMML ModelServer | PMML | v1 | – | v0.10 (KServe) | 3, 4 (PMML4.4.1) | SKLearn PMML |
| SKLearn MLServer | Pickled Model | v2 | v2 | v1.0.0 (MLServer) | 1 | SKLearn Iris V2 |
| SKLearn ModelServer | Pickled Model | v1 | – | v0.10 (KServe) | 1 | SKLearn Iris |
| TFServing | TensorFlow SavedModel | v1 | *tensorflow | 2.6.2 (TFServing Versions) | 2 | TensorFlow flower |
| TorchServe | Eager Model/TorchScript | v1, v2, *torchserve | *torchserve | 0.7.0 (TorchServe) | 1 | TorchServe mnist |
| Triton Inference Server | TensorFlow,TorchScript,ONNX | v2 | v2 | 21.09-py3 (Triton) | 8 (TensoRT), 1, 2 (TensorFlow), 1 (PyTorch), 2 (Triton) Compatibility Matrix | Torchscript cifar |
| XGBoost MLServer | Saved Model | v2 | v2 | v1.0.0 (MLServer) | 1 | XGBoost Iris V2 |
| XGBoost ModelServer | Saved Model | v1 | – | v0.10 (KServe) | 1 | XGBoost Iris |
*tensorflow-除了KServe的预测协议外，tensorflow还实现了自己的预测协议。请参阅：Tensorflow服务预测API文档
*torchserve-PyTorch除了KServe之外，还实现了自己的预引用协议。请参阅：Torchserve gRPC API文档
笔记
提供运行时版本的模型可以用推理服务yaml上的runtimeVersion字段覆盖，我们强烈建议为生产服务设置此字段。
apiVersion: "serving.kserve.io/v1beta1"
kind: "InferenceService"
metadata:
name: "torchscript-cifar"
spec:
predictor:
triton:
storageUri: "gs://kfserving-examples/models/torchscript"
runtimeVersion: 21.08-py3
使用推理服务部署Tensorflow模型
创建HTTP推理服务
创建一个指定框架tensorflow和指向已保存的tensorflow模型的storageUri的推理服务yaml，并将其命名为tensorflow.yaml。
旧架构
apiVersion: "serving.kserve.io/v1beta1"
kind: "InferenceService"
metadata:
name: "flower-sample"
spec:
predictor:
tensorflow:
storageUri: "gs://kfserving-examples/models/tensorflow/flowers"
新架构
apiVersion: "serving.kserve.io/v1beta1"
kind: "InferenceService"
metadata:
name: "flower-sample"
spec:
predictor:
model:
modelFormat:
name: tensorflow
storageUri: "gs://kfserving-examples/models/tensorflow/flowers"
kubectl
kubectl apply -f tensorflow.yaml
期望输出
$ inferenceservice.serving.kserve.io/flower-sample created
等待推理服务处于就绪状态
kubectl get isvc flower-sample
NAME URL READY PREV LATEST PREVROLLEDOUTREVISION LATESTREADYREVISION AGE
flower-sample http://flower-sample.default.example.com True 100 flower-sample-predictor-default-n9zs6 7m15s
运行预测
第一步是确定入口IP和端口，并设置ingress_HOST和ingress_PORT，在这里可以下载推理请求输入文件。
MODEL_NAME=flower-sample
INPUT_PATH=@./input.json
SERVICE_HOSTNAME=$(kubectl get inferenceservice ${MODEL_NAME} -o jsonpath='{.status.url}' | cut -d "/" -f 3)
curl -v -H "Host: ${SERVICE_HOSTNAME}" http://${INGRESS_HOST}:${INGRESS_PORT}/v1/models/$MODEL_NAME:predict -d $INPUT_PATH
期望输出
* Connected to localhost (::1) port 8080 (#0)
> POST /v1/models/tensorflow-sample:predict HTTP/1.1
> Host: tensorflow-sample.default.example.com
> User-Agent: curl/7.73.0
> Accept: */*
> Content-Length: 16201
> Content-Type: application/x-www-form-urlencoded
>
* upload completely sent off: 16201 out of 16201 bytes
* Mark bundle as not supporting multiuse
< HTTP/1.1 200 OK
< content-length: 222
< content-type: application/json
< date: Sun, 31 Jan 2021 01:01:50 GMT
< x-envoy-upstream-service-time: 280
< server: istio-envoy
<
{
"predictions": [
{
"scores": [0.999114931, 9.20987877e-05, 0.000136786213, 0.000337257545, 0.000300532585, 1.84813616e-05],
"prediction": 0,
"key": " 1"
}
]
}
金丝雀Rollout
Canary Rollout是控制推出新模型风险的一个好方法，首先将一小部分流量转移到它身上，然后逐渐增加这个百分比。要运行canary rollout，可以应用指定了canaryTrafficPercent字段的canary.yaml。
旧架构
apiVersion: "serving.kserve.io/v1beta1"
kind: "InferenceService"
metadata:
name: "flower-sample"
spec:
predictor:
canaryTrafficPercent: 20
tensorflow:
storageUri: "gs://kfserving-examples/models/tensorflow/flowers-2"
新架构
apiVersion: "serving.kserve.io/v1beta1"
kind: "InferenceService"
metadata:
name: "flower-sample"
spec:
predictor:
canaryTrafficPercent: 20
model:
modelFormat:
name: tensorflow
storageUri: "gs://kfserving-examples/models/tensorflow/flowers-2"
应用canary.yaml创建金丝雀推理服务。
kubectl
kubectl apply -f canary.yaml
要验证流量分割百分比是否正确应用，可以运行以下命令：
kubectl get isvc flower-sample
NAME URL READY PREV LATEST PREVROLLEDOUTREVISION LATESTREADYREVISION AGE
flower-sample http://flower-sample.default.example.com True 80 20 flower-sample-predictor-default-n9zs6 flower-sample-predictor-default-2kwtr 7m15s
正如您所看到的，流量在上一次推出的修订版和当前最新准备的修订版之间分配，KServe会自动为您跟踪上一次发布的（稳定的）修订版，因此您不需要像v1alpha2那样在推理服务上同时维护default和金丝雀。
创建gRPC推理服务
创建推理服务，它暴露gRPC端口，默认情况下它侦听端口9000。
旧框架
apiVersion: "serving.kserve.io/v1beta1"
kind: "InferenceService"
metadata:
name: "flower-grpc"
spec:
predictor:
tensorflow:
storageUri: "gs://kfserving-examples/models/tensorflow/flowers"
ports:
- containerPort: 9000
name: h2c
protocol: TCP
新框架
apiVersion: "serving.kserve.io/v1beta1"
kind: "InferenceService"
metadata:
name: "flower-grpc"
spec:
predictor:
model:
modelFormat:
name: tensorflow
storageUri: "gs://kfserving-examples/models/tensorflow/flowers"
ports:
- containerPort: 9000
name: h2c
protocol: TCP
应用grp.yaml创建grpc推理服务。
kubectl
kubectl apply -f grpc.yaml
期望输出
$ inferenceservice.serving.kserve.io/flower-grpc created
运行预测
我们使用python gRPC客户端进行预测，因此您需要创建一个python虚拟环境并安装tensorflow服务api。
# The prediction script is written in TensorFlow 1.x
pip install tensorflow-serving-api>=1.14.0,<2.0.0
运行gRPC预测脚本。
MODEL_NAME=flower-grpc
INPUT_PATH=./input.json
SERVICE_HOSTNAME=$(kubectl get inferenceservice ${MODEL_NAME} -o jsonpath='{.status.url}' | cut -d "/" -f 3)
python grpc_client.py --host $INGRESS_HOST --port $INGRESS_PORT --model $MODEL_NAME --hostname $SERVICE_HOSTNAME --input_path $INPUT_PATH
期望输出
outputs {
key: "key"
value {
dtype: DT_STRING
tensor_shape {
dim {
size: 1
}
}
string_val: " 1"
}
}
outputs {
key: "prediction"
value {
dtype: DT_INT64
tensor_shape {
dim {
size: 1
}
}
int64_val: 0
}
}
outputs {
key: "scores"
value {
dtype: DT_FLOAT
tensor_shape {
dim {
size: 1
}
dim {
size: 6
}
}
float_val: 0.9991149306297302
float_val: 9.209887502947822e-05
float_val: 0.00013678647519554943
float_val: 0.0003372581850271672
float_val: 0.0003005331673193723
float_val: 1.848137799242977e-05
}
}
model_spec {
name: "flowers-sample"
version {
value: 1
}
signature_name: "serving_default"
}
使用TorchServe推理服务部署PyTorch模型
在本例中，我们部署了一个经过训练的PyTorch MNIST模型，通过使用TorchServe runtime运行推理服务来预测手写数字，TorchServe运行时是PyTorch模型的默认安装服务运行时。模型可解释性也是一个重要方面，它有助于理解哪些输入特征对特定分类很重要。Captum是一个模型可解释性库。在这个例子中，TorchServe解释端点是用Captum最先进的算法实现的，包括集成的梯度，为用户提供了一种简单的方式来理解哪些特征有助于模型输出。您可以参考Captum教程了解更多示例。
使用模型档案文件和配置创建模型存储
KServe/TorchServe集成需要以下模型商店布局。
├── config
│ ├── config.properties
├── model-store
│ ├── densenet_161.mar
│ ├── mnist.mar
TorchServe提供了一个实用程序，可以将所有模型工件打包到一个TorchServe Model Archive File (MAR)中。将模型工件打包到MAR文件中后，您将上传到模型存储路径下的模型存储。
您可以将模型和相关文件存储在远程存储或本地持久卷上。MNIST模型和相关文件可以从这里获得。
笔记
对于远程存储，您可以选择使用存储在KServe example GCS bucket gs://kfserving examples/models/torchserver/image_classifier上的预构建MNIST MAR文件启动示例，或者使用torch模型归档器生成MAR文件，并根据上述布局在远程存储上创建模型存储。
torch模型归档器--模型名称mnist--1.0版\
--模型文件模型归档器/模型库/mnist/mnist.py\
--序列化文件模型归档器/模型存储/mnist/mnist_cnn.pt\
--处理程序模型归档器/模型存储区/mnist/mnist_handler.py\
对于PVC用户，请参阅模型档案文件生成，以自动生成带有模型和从属文件的MAR文件。
TorchServe使用config.properties文件来存储配置。有关配置文件支持的属性的更多详细信息，请参阅此处。以下是KServe的示例文件：
inference_address=http://0.0.0.0:8085
management_address=http://0.0.0.0:8085
metrics_address=http://0.0.0.0:8082
grpc_inference_port=7070
grpc_management_port=7071
enable_metrics_api=true
metrics_format=prometheus
number_of_netty_threads=4
job_queue_size=10
enable_envvars_config=true
install_py_dep_per_model=true
model_store=/mnt/models/model-store
model_snapshot={"name":"startup.cfg","modelCount":1,"models":{"mnist":{"1.0":{"defaultVersion":true,"marName":"mnist.mar","minWorkers":1,"maxWorkers":5,"batchSize":1,"maxBatchDelay":10,"responseTimeout":120}}}}
KServe/TorchServe集成支持KServe v1/v2 REST协议。在config.properties中，我们需要打开标志enable_envvars_config，以便使用环境变量设置KServe信封。
警告
以前的service_envelope属性已被弃用，并且在config.properties文件中使用标志enable_envvars_config=true来启用在运行时设置服务信封。请求从KServe推理请求格式转换为TorchServe请求格式，并发送到通过本地套接字配置的推理地址。
使用V1 REST协议部署PyTorch模型
创建TorchServe推理服务
当您在新模型规范上指定模型格式pytorch时，KServe默认选择TorchServe运行时。
旧框架
apiVersion: "serving.kserve.io/v1beta1"
kind: "InferenceService"
metadata:
name: "torchserve"
spec:
predictor:
pytorch:
storageUri: gs://kfserving-examples/models/torchserve/image_classifier/v1
新框架
apiVersion: "serving.kserve.io/v1beta1"
kind: "InferenceService"
metadata:
name: "torchserve"
spec:
predictor:
model:
modelFormat:
name: pytorch
storageUri: gs://kfserving-examples/models/torchserve/image_classifier/v1
要在CPU上部署模型，请应用以下torchserver.yaml来创建推理服务。
kubectl
kubectl apply -f torchserve.yaml
旧框架
apiVersion: "serving.kserve.io/v1beta1"
kind: "InferenceService"
metadata:
name: "torchserve"
spec:
predictor:
pytorch:
storageUri: gs://kfserving-examples/models/torchserve/image_classifier/v1
resources:
limits:
memory: 4Gi
nvidia.com/gpu: "1"
新框架
apiVersion: "serving.kserve.io/v1beta1"
kind: "InferenceService"
metadata:
name: "torchserve"
spec:
predictor:
model:
modelFormat:
name: pytorch
storageUri: gs://kfserving-examples/models/torchserve/image_classifier/v1
resources:
limits:
memory: 4Gi
nvidia.com/gpu: "1"
要在GPU上部署模型，请应用gpu.yaml来创建GPU推理服务。
kubectl
kubectl apply -f gpu.yaml
期望输出
$ inferenceservice.serving.kserve.io/torchserve created
模型推理
第一步是确定入口IP和端口并设置INGRESS_HOST 和 INGRESS_PORT
MODEL_NAME=mnist
SERVICE_HOSTNAME=$(kubectl get inferenceservice torchserve -o jsonpath='{.status.url}' | cut -d "/" -f 3)
您可以使用图像转换器将图像转换为基于64字节的数组，其他型号请参阅输入请求。
curl -v -H "Host: ${SERVICE_HOSTNAME}" http://${INGRESS_HOST}:${INGRESS_PORT}/v1/models/${MODEL_NAME}:predict -d @./mnist.json
期望输出
* Trying 52.89.19.61...
* Connected to a881f5a8c676a41edbccdb0a394a80d6-2069247558.us-west-2.elb.amazonaws.com (52.89.19.61) port 80 (#0)
> PUT /v1/models/mnist HTTP/1.1
> Host: torchserve.kserve-test.example.com
> User-Agent: curl/7.47.0
> Accept: */*
> Content-Length: 167
> Expect: 100-continue
>
< HTTP/1.1 100 Continue
* We are completely uploaded and fine
< HTTP/1.1 200 OK
< cache-control: no-cache; no-store, must-revalidate, private
< content-length: 1
< date: Tue, 27 Oct 2020 08:26:19 GMT
< expires: Thu, 01 Jan 1970 00:00:00 UTC
< pragma: no-cache
< x-request-id: b10cfc9f-cd0f-4cda-9c6c-194c2cdaa517
< x-envoy-upstream-service-time: 6
< server: istio-envoy
<
* Connection #0 to host a881f5a8c676a41edbccdb0a394a80d6-2069247558.us-west-2.elb.amazonaws.com left intact
{"predictions": ["2"]}
模型说明
要获取模型说明：
curl -v -H "Host: ${SERVICE_HOSTNAME}" http://${INGRESS_HOST}:${INGRESS_PORT}/v1/models/mnist:explain -d @./mnist.json
期望输出
{"explanations": [[[[0.0005394675730469475, -0.0022280013123036043, -0.003416480100841055, -0.0051329881112415965, -0.009973864160829985, -0.004112560908882716, -0.009223458030656112, -0.0006676354577291628, -0.005249806664413386, -0.0009790519227372953, -0.0026914653993121195, -0.0069470097151383995, -0.00693530415962956, -0.005973878697847718, -0.00425042437288857, 0.0032867281838150977, -0.004297780258633562, -0.005643196661192014, -0.00653025019738562, -0.0047062916121001185, -0.0018656628277792628, -0.0016757477204072532, -0.0010410417081844845, -0.0019093520822156726, -0.004451403461006374, -0.0008552767257773671, -0.0027638888169885267, -0.0], [0.006971297052106784, 0.007316855222185687, 0.012144494329150574, 0.011477799383288441, 0.006846725347670252, 0.01149386176451476, 0.0045351987881190655, 0.007038361889638708, 0.0035855377023272157, 0.003031419502053957, -0.0008611575226775316, -0.0011085224745969223, -0.0050840743637658534, 0.009855491784340777, 0.007220680811043034, 0.011374285598070253, 0.007147725481709019, 0.0037114580912849457, 0.00030763245479291384, 0.0018305492665953394, 0.010106224395114147, 0.012932881164284687, 0.008862892007714321, 0.0070960526615982435, -0.0015931137903787505, 0.0036495747329455906, 0.0002593849391051298, -0.0], [0.006467265785857396, -0.00041793201228071674, 0.004900316089756856, 0.002308395474823997, 0.007859295399592283, 0.003916404948969494, 0.005630750246437249, 0.0043712538044184375, 0.006128530599133763, -0.009446321309831246, -0.014173645867037036, -0.0062988650915794565, -0.011473838941118539, -0.009049151947644047, -0.0007625645864610934, -0.013721416630061238, -0.0005580156670410108, 0.0033404383756480784, -0.006693278798487951, -0.003705084551144756, 0.005100375089529131, 5.5276874714401074e-05, 0.007221745280359063, -0.00573598303916232, -0.006836169033785967, 0.0025401608627538936, 9.303533912921196e-05, -0.0], [0.005914399808621816, 0.00452643561023696, 0.003968242261515448, 0.010422786058967673, 0.007728358107899074, 0.01147115923288383, 0.005683869479056691, 0.011150670502307374, 0.008742555292485278, 0.0032882897575743754, 0.014841138421861584, 0.011741228362482451, 0.0004296862879259221, -0.0035118140680654854, -0.006152254410078331, -0.004925121936901983, -2.3611205202801947e-06, 0.029347073037039074, 0.02901626308947743, 0.023379353021343398, 0.004027157620197582, -0.01677662249919171, -0.013497255736128979, 0.006957482854214602, 0.0018321766800746145, 0.008277034396684563, 0.002733405455464871, -0.0], [0.0049579739156640065, -0.002168016158233997, 0.0020644317321723642, 0.0020912464240293825, 0.004719691119907336, 0.007879231202446626, 0.010594445898145937, 0.006533067778982801, 0.002290214592708113, -0.0036651114968251986, 0.010753227423379443, 0.006402706020466243, -0.047075193909339695, -0.08108259303568185, -0.07646875196692542, -0.1681834845371156, -0.1610307396135756, -0.12010309927453829, -0.016148831320070896, -0.009541525999486027, 0.04575604594761406, 0.031470966329886635, 0.02452149438024385, 0.016594078577569567, 0.012213591301610382, -0.002230875840404426, 0.0036704051254298374, -0.0], [0.006410107592414739, 0.005578283890924384, 0.001977103461731095, 0.008935476507124939, 0.0011305055729953436, 0.0004946313900665659, -0.0040266029554395935, -0.004270765544167256, -0.010832150944943138, -0.01653511868336456, -0.011121302103373972, -0.42038514526905024, -0.22874576003118394, -0.16752936178907055, -0.17021699697722079, -0.09998584936787697, -0.09041117495322142, -0.10230248444795721, -0.15260897522094888, 0.07770835838531896, -0.0813761125123066, 0.027556910053932963, 0.036305965104261866, 0.03407793793894619, 0.01212761779302579, 0.006695133380685627, 0.005331392748588556, -0.0], [0.008342680065996267, -0.00029249776150416367, 0.002782130291086583, 0.0027793744856745373, 0.0020525102690845407, 0.003679269934110004, 0.009373846012918791, -0.0031751745946300403, -0.009042846256743316, 0.0074141593032070775, -0.02796812516561052, -0.593171583786029, -0.4830164472795136, -0.353860128479443, -0.256482708704862, 0.11515586314578445, 0.12700563162828346, 0.0022342450630152204, -0.24673707669992118, -0.012878340813781437, 0.16866821780196756, 0.009739033161051434, -0.000827843726513152, -0.0002137320694585577, -0.004179480126338929, 0.008454049232317358, -0.002767934266266998, -0.0], [0.007070382982749552, 0.005342127805750565, -0.000983984198542354, 0.007910101170274493, 0.001266267696096404, 0.0038575136843053844, 0.006941130321773131, -0.015195182020687892, -0.016954974010578504, -0.031186444096787943, -0.031754626467747966, 0.038918845112017694, 0.06248943950328597, 0.07703301092601872, 0.0438493628024275, -0.0482404449771698, -0.08718650815999045, -0.0014764704694506415, -0.07426336448916614, -0.10378029666564882, 0.008572087846793842, -0.00017173413848283343, 0.010058893270893113, 0.0028410498666004377, 0.002008290211806285, 0.011905375389931099, 0.006071375802943992, -0.0], [0.0076080165949142685, -0.0017127333725310495, 0.00153128150106188, 0.0033391793764531563, 0.005373442509691564, 0.007207746020295443, 0.007422946703693544, -0.00699779191449194, 0.002395328253696969, -0.011682618874195954, -0.012737004464649057, -0.05379966383523857, -0.07174960461749053, -0.03027341304050314, 0.0019411862216381327, -0.0205575129473766, -0.04617091711614171, -0.017655308106959804, -0.009297162816368814, -0.03358572117988279, -0.1626068444778013, -0.015874364762085157, -0.0013736074085577258, -0.014763439328689378, 0.00631805792697278, 0.0021769414283267273, 0.0023061635006792498, -0.0], [0.005569931813561535, 0.004363218328087518, 0.00025609463218383973, 0.009577483244680675, 0.007257755916229399, 0.00976284778532342, -0.006388840235419147, -0.009017880790555707, -0.015308709334434867, -0.016743935775597355, -0.04372596546189275, -0.03523469356755156, -0.017257810114846107, 0.011960489902313411, 0.01529079831828911, -0.020076559119468443, -0.042792547669901516, -0.0029492027218867116, -0.011109560582516062, -0.12985858077848939, -0.2262858575494602, -0.003391725540087574, -0.03063368684328981, -0.01353486587575121, 0.0011140822443932317, 0.006583451102528798, 0.005667533945285076, -0.0], [0.004056272267155598, -0.0006394041203204911, 0.004664893926197093, 0.010593032387298614, 0.014750931538689989, 0.015428721146282149, 0.012167820222401367, 0.017604752451202518, 0.01038886849969188, 0.020544326931163263, -0.0004206566917812794, -0.0037463581359232674, -0.0024656693040735075, 0.0026061897697624353, -0.05186055271869177, -0.09158655048397382, 0.022976389912563913, -0.19851635458461808, -0.11801281807622972, -0.29127727790584423, -0.017138655663803876, -0.04395515676468641, -0.019241432506341576, 0.0011342298743447392, 0.0030625771422964584, -0.0002867924892991192, -0.0017908808807543712, -0.0], [0.0030114260660488892, 0.0020246448273580006, -0.003293361220376816, 0.0036965043883218584, 0.00013185761728146236, -0.004355610866966878, -0.006432601921104354, -0.004148701459814858, 0.005974553907915845, -0.0001399233607281906, 0.010392944122965082, 0.015693249298693028, 0.0459528427528407, -0.013921539948093455, -0.06615556518538708, 0.02921438991320325, -0.16345220625101778, -0.002130491295590408, -0.11449749664916867, -0.030980255589300607, -0.04804122537359171, -0.05144994776295644, 0.005122827412776085, 0.006464862173908011, 0.008624278272940246, 0.0037316228508156427, 0.0036947794337026706, -0.0], [0.0038173843228389405, -0.0017091931226819494, -0.0030871869816778068, 0.002115642501535999, -0.006926441921580917, -0.003023077828426468, -0.014451359520861637, -0.0020793048380231397, -0.010948003939342523, -0.0014460716966395166, -0.01656990336897737, 0.003052317148320358, -0.0026729564809943513, -0.06360067057346147, 0.07780985635080599, -0.1436689936630281, -0.040817177623437874, -0.04373367754296477, -0.18337299150349698, 0.025295182977407064, -0.03874921104331938, -0.002353901742617205, 0.011772560401335033, 0.012480994515707569, 0.006498422579824301, 0.00632320984076023, 0.003407169765754805, -0.0], [0.00944355257990139, 0.009242583578688485, 0.005069860444386138, 0.012666191449103024, 0.00941789912565746, 0.004720427012836104, 0.007597687789204113, 0.008679266528089945, 0.00889322771021875, -0.0008577904940828809, 0.0022973860384607604, 0.025328230809207493, -0.09908781123080951, -0.07836626399832172, -0.1546141264726177, -0.2582207272050766, -0.2297524599578219, -0.29561835103416967, 0.12048787956671528, -0.06279365699861471, -0.03832012404275233, 0.022910264999199934, 0.005803508497672737, -0.003858461926053348, 0.0039451232171312765, 0.003858476747495933, 0.0013034515558609956, -0.0], [0.009725756015628606, -0.0004001101998876524, 0.006490722835571152, 0.00800808023631959, 0.0065880711806331265, -0.0010264326176194034, -0.0018914305972878344, -0.008822522194658438, -0.016650520788128117, -0.03254382594389507, -0.014795713101569494, -0.05826499837818885, -0.05165369567511702, -0.13384277337594377, -0.22572641373340493, -0.21584739544668635, -0.2366836351939208, 0.14937824076489659, -0.08127414932170171, -0.06720440139736879, -0.0038552732903526744, 0.0107597891707803, -5.67453590118174e-05, 0.0020161340511396244, -0.000783322694907436, -0.0006397207517995289, -0.005291639205010064, -0.0], [0.008627543242777584, 0.007700097300051849, 0.0020430960246806138, 0.012949015733198586, 0.008428709579953574, 0.001358177022953576, 0.00421863939925833, 0.002657580000868709, -0.007339431957237175, 0.02008439775442315, -0.0033717631758033114, -0.05176633249899187, -0.013790328758662772, -0.39102366157050594, -0.167341447585844, -0.04813367828213947, 0.1367781582239039, -0.04672809260566293, -0.03237784669978756, 0.03218068777925178, 0.02415063765016493, -0.017849899351200002, -0.002975675228088795, -0.004819438014786686, 0.005106898651831245, 0.0024278620704227456, 6.784303333368138e-05, -0.0], [0.009644258527009343, -0.001331907219439711, -0.0014639718434477777, 0.008481926798958248, 0.010278031715467508, 0.003625808326891529, -0.01121188617599796, -0.0010634587872994379, -0.0002603820881968461, -0.017985648016990465, -0.06446652745470374, 0.07726063173046191, -0.24739929795334742, -0.2701855018480216, -0.08888614776216278, 0.1373325760136816, -0.02316068912438066, -0.042164834956711514, 0.0009266091344106458, 0.03141872420427644, 0.011587728430225652, 0.0004755143243520787, 0.005860642609620605, 0.008979633931394438, 0.005061734169974005, 0.003932710387086098, 0.0015489986106803626, -0.0], [0.010998736164377534, 0.009378969800902604, 0.00030577045264713074, 0.0159329353530375, 0.014849508018911006, -0.0026513365659554225, 0.002923303082126996, 0.01917908707828847, -0.02338288107991566, -0.05706674679291175, 0.009526265752669624, -0.19945255386401284, -0.10725519695909647, -0.3222906835083537, -0.03857038318412844, -0.013279804965996065, -0.046626023244262085, -0.029299060237210447, -0.043269580558906555, -0.03768510002290657, -0.02255977771908117, -0.02632588166863199, -0.014417349488098566, -0.003077271951572957, -0.0004973277708010661, 0.0003475839139671271, -0.0014522783025903258, -0.0], [0.012215315671616316, -0.001693194176229889, 0.011365785434529038, 0.0036964574178487792, -0.010126738168635003, -0.025554378647710443, 0.006538003839811914, -0.03181759044467965, -0.016424751042854728, 0.06177539736110035, -0.43801735323216856, -0.29991040815937386, -0.2516019795363623, 0.037789523540809, -0.010948746374759491, -0.0633901687126727, -0.005976006160777705, 0.006035133605976937, -0.04961632526071937, -0.04142116972831476, -0.07558952727782252, -0.04165176179187153, -0.02021603856619006, -0.0027365663096057032, -0.011145473712733575, 0.0003566937349350848, -0.00546472985268321, -0.0], [0.008009386447317503, 0.006831207743885825, 0.0051306149795546365, 0.016239014770865052, 0.020925441734273218, 0.028344800173195076, -0.004805080609285047, -0.01880521614501033, -0.1272329010865855, -0.39835936819190537, -0.09113694760349819, -0.04061591094832608, -0.12677021961235907, 0.015567707226741051, -0.005615051546243333, -0.06454044862001587, 0.0195457674752272, -0.04219686517155871, -0.08060569979524296, 0.027234494361702787, -0.009152881336047056, -0.030865118003992217, -0.005770311060090559, 0.002905833371986098, 5.606663556872091e-05, 0.003209538083839772, -0.0018588810743365345, -0.0], [0.007587008852984699, -0.0021213639853557625, 0.0007709558092903736, 0.013883256128746423, 0.017328713012428214, 0.03645357525636198, -0.04043993335238427, 0.05730125171252314, -0.2563293727512057, -0.11438826083879326, 0.02662382809034687, 0.03525271352483709, 0.04745678120172762, 0.0336360484090392, -0.002916635707204059, -0.17950855098650784, -0.44161773297052964, -0.4512180227831197, -0.4940283106297913, -0.1970108671285798, 0.04344323143078066, -0.012005120444897523, 0.00987576109166055, -0.0018336757466252476, 0.0004913959502151706, -0.0005409724034216215, -0.005039223900868212, -0.0], [0.00637876531169957, 0.005189469227685454, 0.0007676355246000376, 0.018378100865097655, 0.015739815031394887, -0.035524983116512455, 0.03781006978038308, 0.28859052096740495, 0.0726464110153121, -0.026768468497420147, 0.06278766200288134, 0.17897045813699355, -0.13780371920803108, -0.14176458123649577, -0.1733103177731656, -0.3106508869296763, 0.04788355140275794, 0.04235327890285105, -0.031266625292514394, -0.016263819217960652, -0.031388328800811355, -0.01791363975905968, -0.012025067979443894, 0.008335083985905805, -0.0014386677797296231, 0.0055376544652972854, 0.002241522815466253, -0.0], [0.007455256326741617, -0.0009475207572210404, 0.0020288385162615286, 0.015399640135796092, 0.021133843188103074, -0.019846405097622234, -0.003162485751163173, -0.14199005055318842, -0.044200898667146035, -0.013395459413208084, 0.11019680479230103, -0.014057216041764874, -0.12553853334447865, -0.05992513534766256, 0.06467942189539834, 0.08866056095907732, -0.1451321508061849, -0.07382491447758655, -0.046961739981080476, 0.0008943713493160624, 0.03231044103656507, 0.00036034241706501196, -0.011387669277619417, -0.00014602449257226195, -0.0021863729003374116, 0.0018817840156005856, 0.0037909804578166286, -0.0], [0.006511855618626698, 0.006236866054439829, -0.001440571166157676, 0.012795776609942026, 0.011530545030403624, 0.03495489377257363, 0.04792403136095304, 0.049378583599065225, 0.03296101702085617, -0.0005351385876652296, 0.017744115897640366, 0.0011656622496764954, 0.0232845869823761, -0.0561191397060232, -0.02854070511118366, -0.028614174047247348, -0.007763531086362863, 0.01823079560098924, 0.021961392405283622, -0.009666681805706179, 0.009547046884328725, -0.008729943263791338, 0.006408909680578429, 0.009794327096359952, -0.0025825219195515304, 0.007063559189211571, 0.007867244119267047, -0.0], [0.007936663546039311, -0.00010710180170593153, 0.002716512705673228, 0.0038633557307721487, -0.0014877316616940372, -0.0004788143065635909, 0.012508842248031202, 0.0045381104608414645, -0.010650910516128294, -0.013785341529644855, -0.034287643221318206, -0.022152707546335495, -0.047056481347685974, -0.032166744564720455, -0.021551611335278546, -0.002174962503376043, 0.024344287130424306, 0.015579272560525105, 0.010958169741952194, -0.010607232913436921, -0.005548369726118836, -0.0014630046444242706, 0.013144180105016433, 0.0031349366359021916, 0.0010984887428255974, 0.005426941473328394, 0.006566511860044785, -0.0], [0.0005529184874606495, 0.00026139355020588705, -0.002887623443531047, 0.0013988462990850632, 0.00203365139495493, -0.007276926701775218, -0.004010419939595932, 0.017521952161185662, 0.0006996977433557911, 0.02083134683611201, 0.013690533534289498, -0.005466724359976675, -0.008857712321334327, 0.017408578822635818, 0.0076439343049154425, 0.0017861314923539985, 0.007465865707523924, 0.008034420825988495, 0.003976298558337994, 0.00411970637898539, -0.004572592545819698, 0.0029563907011979935, -0.0006382227820088148, 0.0015153753877889707, -0.0052626601797995595, 0.0025664706985019416, 0.005161751034260073, -0.0], [0.0009424280561998445, -0.0012942360298110595, 0.0011900868416523343, 0.000984424113178899, 0.0020988269382781564, -0.005870080062890889, -0.004950484744457169, 0.003117643454332697, -0.002509563565777083, 0.005831604884101081, 0.009531085216183116, 0.010030206821909806, 0.005858190171099734, 4.9344529936340524e-05, -0.004027895832421331, 0.0025436439920587606, 0.00531153867563076, 0.00495942692369508, 0.009215148318606382, 0.00010011928317543458, 0.0060051362999805355, -0.0008195376963202741, 0.0041728603512658224, -0.0017597169567888774, -0.0010577007775543158, 0.00046033327178068433, -0.0007674196306044449, -0.0], [-0.0, -0.0, 0.0013386963856532302, 0.00035183178922260837, 0.0030610334903526204, 8.951834979315781e-05, 0.0023676793550483524, -0.0002900551076915047, -0.00207019445286608, -7.61697478482574e-05, 0.0012150086715244216, 0.009831239281792168, 0.003479667642621962, 0.0070584324334114525, 0.004161851261339585, 0.0026146296354490665, -9.194746959222099e-05, 0.0013583866966571571, 0.0016821551239318913, -0.0, -0.0, -0.0, -0.0, -0.0, -0.0, -0.0, -0.0, -0.0]]]]}
使用V2 REST协议部署PyTorch模型
创建推理服务
当您在新的模型规范中指定模型格式pytorch并启用KServe v1推理协议时，KServe默认选择TorchServe运行时。要启用v2推理协议，请指定值为v2的protocolVersion字段。
旧框架
apiVersion: "serving.kserve.io/v1beta1"
kind: "InferenceService"
metadata:
name: "torchserve-mnist-v2"
spec:
predictor:
pytorch:
protocolVersion: v2
storageUri: gs://kfserving-examples/models/torchserve/image_classifier/v2
新框架
apiVersion: "serving.kserve.io/v1beta1"
kind: "InferenceService"
metadata:
name: "torchserve-mnist-v2"
spec:
predictor:
model:
modelFormat:
name: pytorch
protocolVersion: v2
storageUri: gs://kfserving-examples/models/torchserve/image_classifier/v2
要在CPU上部署模型，请应用mnist_v2.yaml来创建推理服务。
kubectl
kubectl apply -f mnist_v2.yaml
期望输出
$ inferenceservice.serving.kserve.io/torchserve-mnist-v2 created
模型推理
第一步是确定入口IP和端口，并设置INGRESS_HOST和INGRESS_PORT。
MODEL_NAME=mnist
SERVICE_HOSTNAME=$(kubectl get inferenceservice torchserve-mnist-v2 -o jsonpath='{.status.url}' | cut -d "/" -f 3)
您可以使用v2协议发送字节数组和张量，对于字节数组，使用图像转换器将图像转换为字节数组输入。在这里，我们使用mnist_v2_bytes.json文件来运行一个示例推理。
curl -v -H "Host: ${SERVICE_HOSTNAME}" http://${INGRESS_HOST}:${INGRESS_PORT}/v2/models/${MODEL_NAME}/infer -d @./mnist_v2_bytes.json
期望输出
{"id": "d3b15cad-50a2-4eaf-80ce-8b0a428bd298", "model_name": "mnist", "model_version": "1.0", "outputs": [{"name": "predict", "shape": [], "datatype": "INT64", "data": [1]}]}
对于张量输入，使用张量图像转换器将图像转换为张量输入，这里我们使用mnist_v2.json文件来运行示例推理。
curl -v -H "Host: ${SERVICE_HOSTNAME}" http://${INGRESS_HOST}:${INGRESS_PORT}/v2/models/${MODEL_NAME}/infer -d @./mnist_v2.json
期望输出
{"id": "2266ec1e-f600-40af-97b5-7429b8195a80", "model_name": "mnist", "model_version": "1.0", "outputs": [{"name": "predict", "shape": [], "datatype": "INT64", "data": [1]}]}
模型说明
要使用v2解释端点获取模型解释，请执行以下操作：
MODEL_NAME=mnist
curl -v -H "Host: ${SERVICE_HOSTNAME}" http://${INGRESS_HOST}:${INGRESS_PORT}/v2/models/mnist/explain -d @./mnist_v2.json
期望输出
{"id": "d3b15cad-50a2-4eaf-80ce-8b0a428bd298", "model_name": "mnist", "model_version": "1.0", "outputs": [{"name": "explain", "shape": [1, 28, 28], "datatype": "FP64", "data": [-0.0, -0.0, -0.0, -0.0, -0.0, -0.0, -0.0, -0.0, -0.0, -0.0, -0.0, -0.0, -0.0, -0.0, -0.0, -0.0, 0.0, -0.0, -0.0, 0.0, -0.0, 0.0, -0.0, -0.0, -0.0, -0.0, -0.0, 0.0, -0.0, -0.0, -0.0, -0.0, -0.0, -0.0, -0.0, 0.0, -0.0, 0.0, -0.0, -0.0, -0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, -0.0, -0.0, 0.0, 0.0, -0.0, 0.0, -0.0, -0.0, -0.0, -0.0, -0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, -0.0, -0.0, -0.0, 0.0, -0.0, 0.0, 0.0, 0.0, -0.0, -0.0, -0.0, 0.0, -0.0, 0.0, -0.0, -0.0, -0.0, -0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, -0.0, -0.0, 0.0, 0.0, -0.0, -0.0, -0.0, -0.0, -0.0, 0.0, 0.0, -0.0, -0.0, -0.0, 0.0, 0.0, 0.0, -0.0, -0.0, -0.0, -0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, -0.0, -0.0040547528781588035, -0.00022612877200043775, -0.0001273413606783097, 0.005648369508785856, 0.008904784451506994, 0.0026385365879584796, 0.0026802458602499875, -0.002657801604900743, -0.0, -0.0, -0.0, -0.0, -0.0, -0.0, 0.0, -0.0, -0.0, -0.0, -0.0, -0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.00024465772895309256, 0.0008218449738666515, 0.015285917610467934, 0.007512832227517626, 0.007094984753782517, 0.003405668751094489, -0.0020919252360163056, -0.00078002938659872, 0.02299587777864007, 0.01900432942654754, -0.001252955497754338, -0.0014666116894338772, -0.0, -0.0, -0.0, 0.0, 0.0, 0.0, 0.0, -0.0, -0.0, -0.0, -0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.005298396384926053, -0.0007901605067151054, 0.0039060659788228954, 0.02317408211645009, 0.017237917554858186, 0.010867034286601965, 0.003001563092717309, 0.00622421762838887, 0.006120712336480808, 0.016736329175541464, 0.005674718838256385, 0.004344134814439431, -0.001232842177319105, -0.0, -0.0, -0.0, 0.0, 0.0, -0.0, -0.0, -0.0, 0.0, 0.0, 0.0, 0.0, 0.0, -0.0, 0.0006867353660007012, 0.00977289933298656, -0.003875493166540815, 0.0017986937404117591, 0.0013075440157543057, -0.0024510980461748236, -0.0008806773426546923, -0.0, -0.0, -0.00014277890422995419, -0.009322313284511257, 0.020608317953885236, 0.004351394739722548, -0.0007875565409186222, -0.0009075897751127677, -0.0, -0.0, 0.0, 0.0, 0.0, -0.0, -0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.00022247237111456804, -0.0007829031603535926, 0.002666369539125161, 0.000973336852105775, 0.0, -0.0, 0.0, 0.0, 0.0, 0.0, -0.0, 0.000432321003928822, 0.023657172129172684, 0.010694844898905204, -0.002375952975746018, -0.0, -0.0, 0.0, 0.0, -0.0, -0.0, -0.0, 0.0, 0.0, 0.0, 0.0, 0.0, -0.0, -0.0020747972047037, -0.002320101258915877, -0.0012899205783904548, 0.0, 0.0, 0.0, -0.0, -0.0, -0.0, -0.0, -0.0, -0.0, 0.007629679655402933, 0.01044862724376463, 0.00025032878924736025, -0.0, -0.0, 0.0, 0.0, 0.0, -0.0, -0.0, 0.0, 0.0, 0.0, 0.0, -0.0, -0.0, -0.00037708370104137974, -0.005156369275302328, 0.0012477582442296628, 0.0, 0.0, 0.0, 0.0, 0.0, -0.0, -0.0, 0.0, -0.0, -4.442516083381132e-05, 0.01024804634283815, 0.0009971135240970147, -0.0, -0.0, 0.0, 0.0, 0.0, -0.0, -0.0, 0.0, -0.0, 0.0, 0.0, -0.0, 0.0004501048968956462, -0.0019630535686311007, -0.0006664793297549408, 0.0020157403539278907, 0.0, 0.0, -0.0, -0.0, -0.0, -0.0, -0.0, 0.0, -0.0, -0.0022144569383238466, 0.008361583574785395, 0.00314019428604999, -0.0, -0.0, 0.0, 0.0, 0.0, -0.0, -0.0, -0.0, -0.0, -0.0, -0.0, -0.0, -0.0028943544591141838, -0.0031301383432286406, 0.002113252872926688, 0.0, 0.0, 0.0, 0.0, 0.0, -0.0, -0.0, 0.0, -0.0, -0.0, -0.0010321050605717045, 0.008905753926369048, 0.002846438277738756, -0.0, -0.0, 0.0, 0.0, 0.0, -0.0, -0.0, -0.0, -0.0, -0.0, -0.0, 0.0, -0.005305288883499087, -0.00192711009725932, 0.0012090042768467344, 0.0, 0.0, 0.0, -0.0, -0.0, -0.0, 0.0, 0.0, 0.0, -0.0, -0.0011945156500241256, 0.005654442715832439, 0.0020132075345016807, -0.0, -0.0, 0.0, 0.0, 0.0, -0.0, -0.0, -0.0, -0.0, -0.0, -0.0, 0.0, -0.0014689356969061985, 0.0010743412638183228, 0.0, 0.0, 0.0, 0.0, -0.0, -0.0, -0.0, -0.0, 0.0, -0.0, -0.0, -0.0017047980586912376, 0.00290660517425009, -0.0007805869640505143, -0.0, -0.0, 0.0, 0.0, 0.0, -0.0, -0.0, -0.0, 0.0, -0.0, -0.0, 5.541* Connection #0 to host localhost left intact
725422148614e-05, 0.0014516114512869852, 0.0002827701966546988, 0.0, 0.0, 0.0, -0.0, -0.0, -0.0, 0.0, 0.0, 0.0, 0.0, -0.0, -0.0014401407633627265, 0.0023812497776698745, 0.002146825301700187, -0.0, -0.0, 0.0, -0.0, 0.0, -0.0, -0.0, -0.0, -0.0, -0.0, -0.0, 0.0011500529125940918, 0.0002865015572973405, 0.0029798151042282686, 0.0, 0.0, 0.0, -0.0, -0.0, -0.0, 0.0, 0.0, 0.0, -0.0, -0.0, -0.0017750295500283872, 0.0008339859126060243, -0.00377073933577687, -0.0, -0.0, 0.0, 0.0, 0.0, -0.0, -0.0, -0.0, -0.0, 0.0, 0.0, -0.0006093176894575109, -0.00046905787892409935, 0.0034053218511795034, 0.0, 0.0, 0.0, 0.0, -0.0, -0.0, -0.0, 0.0, -0.0, -0.0, -0.0007450011768391558, 0.001298767372877851, -0.008499247640112315, -6.145166131400234e-05, -0.0, -0.0, -0.0, 0.0, 0.0, -0.0, -0.0, -0.0, 0.0, -0.0, 0.0, 0.0011809726042792137, -0.001838476328106708, 0.00541110661116898, 0.0, 0.0, 0.0, 0.0, -0.0, -0.0, -0.0, -0.0, 0.0, -0.002139234224224006, 0.0003259163407641124, -0.005276118873855287, -0.001950984007438105, -9.545670742026532e-07, 0.0, -0.0, 0.0, 0.0, 0.0, -0.0, -0.0, -0.0, -0.0, 0.0, 0.0, 0.0007772404228681039, -0.0001517956264720738, 0.0064814848131711815, -0.0, 0.0, 0.0, -0.0, -0.0, -0.0, -0.0, -0.0, 8.098064985902114e-05, -0.00249042660692983, -0.0020718619200672302, -5.341117902942147e-05, -0.00045564724429915073, 0.0, -0.0, -0.0, 0.0, 0.0, 0.0, -0.0, -0.0, -0.0, -0.0, 0.0, 0.0, 0.0, 0.0022750983476959733, 0.0017164060958460778, 0.0003221344707738082, -0.0, -0.0, -0.0, -0.0, -0.0, -0.0015560282678744543, 9.107238495871273e-05, 0.0008772841497928399, 0.0006502978626355868, -0.004128780767525651, 0.0006030386900152659, 0.0, -0.0, 0.0, -0.0, -0.0, 0.0, 0.0, -0.0, -0.0, 0.0, -0.0, -0.0, 0.0, 0.0, 0.001395995791096219, 0.0026791526689584344, 0.0023995008266391488, -0.0004496096312746451, 0.003101832450753724, 0.007494536066960778, 0.0028641187148287965, -0.0030525907182629075, 0.003420222396518567, 0.0014924018363498125, -0.0009357388301326025, 0.0007856228933169799, -0.0018433973914981437, 1.6031856831240914e-05, 0.0, 0.0, -0.0, -0.0, 0.0, 0.0, -0.0, -0.0, -0.0, -0.0, -0.0, -0.0, -0.0, 0.0, -0.0006999018502034005, 0.004382250870697946, -0.0035419313267119365, -0.0028896748092595375, -0.00048734542493666705, -0.0060873452419295, 0.000388224990424471, 0.002533641537585585, -0.004352836563597573, -0.0006079418766875505, -0.0038101334053377753, -0.000828441340357984, 0.0, -0.0, 0.0, 0.0, -0.0, 0.0, 0.0, 0.0, -0.0, -0.0, -0.0, 0.0, -0.0, -0.0, -0.0, -0.0, -0.0, 0.0010901530866342661, -0.013135008038845744, 0.0004734518707654666, 0.002050423283568135, -0.006609451922460863, 0.0023647861820124366, 0.0046789204256194, -0.0018122527412311837, 0.002137538353955849, 0.0, -0.0, -0.0, 0.0, 0.0, -0.0, -0.0, -0.0, 0.0, 0.0, 0.0, -0.0, -0.0, 0.0, -0.0, -0.0, -0.0, -0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, -0.0, -0.0, -0.0, 0.0, -0.0, -0.0, 0.0, 0.0, 0.0, 0.0, 0.0, -0.0, -0.0, -0.0, 0.0, 0.0, -0.0, -0.0, 0.0, -0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, -0.0, -0.0, 0.0, -0.0, -0.0, -0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, -0.0, -0.0, -0.0, 0.0, 0.0, -0.0, -0.0, -0.0, 0.0, 0.0, 0.0, 0.0, -0.0, -0.0, -0.0, -0.0, -0.0, -0.0, -0.0, -0.0, 0.0, 0.0, -0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, -0.0, -0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, -0.0, 0.0, -0.0, -0.0, -0.0, -0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0]}]}
自动缩放
主要的无服务器推理功能之一是自动缩放与传入工作负载匹配的推理服务的副本。KServe默认启用Knative Pod Autoscaler，它可以监视流量并根据配置的指标进行上下缩放。
Knative Autoscaler
KServe支持Knative Pod Autoscaler（KPA）和Kubernetes的Horizontal Pod Autocaler（HPA）的实现。下面列出了每个自动缩放器的功能和限制。
笔记
如果你想使用Kubernetes Horizontal Pod Autoscaler（HPA），你必须安装HPA扩展
Knative Pod Autoscaler（KPA）
- Knative Serving核心的一部分，安装Knative Serving后默认启用。
- 支持缩放到零功能。
- 不支持基于CPU的自动缩放。
Horizontal Pod Autoscaler（HPA）
- 不是Knative Serving核心的一部分，必须在安装Knative Serving后启用。
- 不支持缩放到零功能。
- 支持基于CPU的自动缩放。
创建具有并发目标的推理服务
硬/软自动缩放限制
您可以使用注释autoscaling.knactive.dev/target配置推理服务以获得软限制。软限制是一个有针对性的限制，而不是严格执行的限制，特别是如果突然出现请求突发，可能会超过这个值。
旧框架
apiVersion: "serving.kserve.io/v1beta1"
kind: "InferenceService"
metadata:
name: "torchserve"
annotations:
autoscaling.knative.dev/target: "10"
spec:
predictor:
pytorch:
storageUri: "gs://kfserving-examples/models/torchserve/image_classifier/v1"
新框架
apiVersion: "serving.kserve.io/v1beta1"
kind: "InferenceService"
metadata:
name: "torchserve"
annotations:
autoscaling.knative.dev/target: "10"
spec:
predictor:
model:
modelFormat:
name: pytorch
storageUri: "gs://kfserving-examples/models/torchserve/image_classifier/v1"
您也可以配置具有字段containerConcurrency和硬限制的推理服务。硬性限制是一个强制的上限。如果并发达到硬限制，多余的请求将被缓冲，并且必须等到有足够的空闲容量来执行请求。
旧框架
apiVersion: "serving.kserve.io/v1beta1"
kind: "InferenceService"
metadata:
name: "torchserve"
spec:
predictor:
containerConcurrency: 10
pytorch:
storageUri: "gs://kfserving-examples/models/torchserve/image_classifier/v1"
新框架
apiVersion: "serving.kserve.io/v1beta1"
kind: "InferenceService"
metadata:
name: "torchserve"
spec:
predictor:
containerConcurrency: 10
model:
modelFormat:
name: pytorch
storageUri: "gs://kfserving-examples/models/torchserve/image_classifier/v1"
在指定缩放目标的软限制或硬限制后，现在可以使用autoscaling.yaml部署推理服务。
kubectl
kubectl apply -f autoscaling.yaml
期望输出
$ inferenceservice.serving.kserve.io/torchserve created
使用并发请求运行推理
第一步是安装hey load generator，然后将并发请求发送到推理服务。
go get -u github.com/rakyll/hey
MODEL_NAME=mnist
SERVICE_HOSTNAME=$(kubectl get inferenceservice torchserve -o jsonpath='{.status.url}' | cut -d "/" -f 3)
hey -m POST -z 30s -D ./mnist.json -host ${SERVICE_HOSTNAME} http://${INGRESS_HOST}:${INGRESS_PORT}/v1/models/${MODEL_NAME}:predict
Check Pod Autoscaling
hey默认情况下会同时生成50个请求，因此您可以看到，当容器并发目标设置为10时，推理服务会扩展到5个pod。
期望输出
NAME READY STATUS RESTARTS AGE
torchserve-predictor-default-cj2d8-deployment-69444c9c74-67qwb 2/2 Terminating 0 103s
torchserve-predictor-default-cj2d8-deployment-69444c9c74-nnxk8 2/2 Terminating 0 95s
torchserve-predictor-default-cj2d8-deployment-69444c9c74-rq8jq 2/2 Running 0 50m
torchserve-predictor-default-cj2d8-deployment-69444c9c74-tsrwr 2/2 Running 0 113s
torchserve-predictor-default-cj2d8-deployment-69444c9c74-vvpjl 2/2 Running 0 109s
torchserve-predictor-default-cj2d8-deployment-69444c9c74-xvn7t 2/2 Terminating 0 103s
金丝雀发布
金丝雀发布是一种部署策略，当您向生产流量的一小部分发布新版本的模型时。
使用Canary模型创建推理服务
在上面的实验之后，现在让我们看看如何在不将全部流量默认移动到新模型的情况下推出新模型。
旧框架
apiVersion: "serving.kserve.io/v1beta1"
kind: "InferenceService"
metadata:
name: "torchserve"
annotations:
serving.kserve.io/enable-tag-routing: "true"
spec:
predictor:
canaryTrafficPercent: 20
pytorch:
storageUri: "gs://kfserving-examples/models/torchserve/image_classifier/v2"
新框架
apiVersion: "serving.kserve.io/v1beta1"
kind: "InferenceService"
metadata:
name: "torchserve"
annotations:
serving.kserve.io/enable-tag-routing: "true"
spec:
predictor:
canaryTrafficPercent: 20
model:
modelFormat:
name: pytorch
storageUri: "gs://kfserving-examples/models/torchserve/image_classifier/v2"
在本例中，我们使用canaryTrafficPercent字段将storageUri更改为v2版本，然后应用canary.yaml。
kubectl
kubectl apply -f canary.yaml
期望输出
kubectl get revisions -l serving.kserve.io/inferenceservice=torchserve
NAME CONFIG NAME K8S SERVICE NAME GENERATION READY REASON ACTUAL REPLICAS DESIRED REPLICAS
torchserve-predictor-default-00001 torchserve-predictor-default 1 True 1 1
torchserve-predictor-default-00002 torchserve-predictor-default 2 True 1 1
kubectl get pods -l serving.kserve.io/inferenceservice=torchserve
NAME READY STATUS RESTARTS AGE
torchserve-predictor-default-00001-deployment-7d99979c99-p49gk 2/2 Running 0 28m
torchserve-predictor-default-00002-deployment-c6fcc65dd-rjknq 2/2 Running 0 3m37s
检查流量状况
金丝雀模型推出后，流量应在金丝雀模型修订版和“稳定”修订版之间进行分配，后者以100%的流量推出，现在从推理服务流量状态检查流量分配：
kubectl get isvc torchserve -ojsonpath='{.status.components}'
期望输出
{
"predictor": {
"address": {
"url": "http://torchserve-predictor-default.default.svc.cluster.local"
},
"latestCreatedRevision": "torchserve-predictor-default-00002",
"latestReadyRevision": "torchserve-predictor-default-00002",
"latestRolledoutRevision": "torchserve-predictor-default-00001",
"traffic": [
{
"latestRevision": true,
"percent": 20,
"revisionName": "torchserve-predictor-default-00002",
"tag": "latest",
"url": "http://latest-torchserve-predictor-default.default.example.com"
},
{
"latestRevision": false,
"percent": 80,
"revisionName": "torchserve-predictor-default-00001",
"tag": "prev",
"url": "http://prev-torchserve-predictor-default.default.example.com"
}
],
"url": "http://torchserve-predictor-default.default.example.com"
}
}
“流量”推出
向InferenceService运行以下curl请求几次，您可以看到请求被发送到具有20/80拆分的两个修订版。
MODEL_NAME=mnist
SERVICE_HOSTNAME=$(kubectl get inferenceservice torchserve -o jsonpath='{.status.url}' | cut -d "/" -f 3)
for i in {1..10}; do curl -H "Host: ${SERVICE_HOSTNAME}" http://${INGRESS_HOST}:${INGRESS_PORT}/v1/models/${MODEL_NAME}:predict -d @./mnist.json; done
期望输出
{"predictions": [2]}Handling connection for 8080
{"predictions": [2]}Handling connection for 8080
{"predictions": [2]}Handling connection for 8080
500: Internal Server Error<<span class="token operator">/</span>title><body>500: Internal Server Error<<span class="token operator">/</span>body><<span class="token operator">/</span>html>Handling connection <span class="token keyword">for</span> 8080
<html><title>500: Internal Server Error<<span class="token operator">/</span>title><body>500: Internal Server Error<<span class="token operator">/</span>body><<span class="token operator">/</span>html>Handling connection <span class="token keyword">for</span> 8080
<span class="token punctuation">{</span><span class="token string">"predictions"</span>: <span class="token punctuation">[</span>2<span class="token punctuation">]</span><span class="token punctuation">}</span>Handling connection <span class="token keyword">for</span> 8080
<span class="token punctuation">{</span><span class="token string">"predictions"</span>: <span class="token punctuation">[</span>2<span class="token punctuation">]</span><span class="token punctuation">}</span>Handling connection <span class="token keyword">for</span> 8080
<span class="token punctuation">{</span><span class="token string">"predictions"</span>: <span class="token punctuation">[</span>2<span class="token punctuation">]</span><span class="token punctuation">}</span>Handling connection <span class="token keyword">for</span> 8080
<span class="token punctuation">{</span><span class="token string">"predictions"</span>: <span class="token punctuation">[</span>2<span class="token punctuation">]</span><span class="token punctuation">}</span>Handling connection <span class="token keyword">for</span> 8080
</code></pre>
<p>您可以注意到，当请求命中canary修订版时，它会失败，这是因为新修订版需要v2推理输入mnist_v2.json，这是一个突破性的更改，此外，流量会根据指定的流量百分比在两个修订版之间随机分配。在这种情况下，您应该使用0 canaryTrafficPercent来推出金丝雀模型，并在将全部流量移动到新模型之前使用最新标记的url来测试金丝雀模型。<br> <code>kubectl</code></p>
<pre><code class="prism language-powershell">kubectl patch isvc torchserve <span class="token operator">--</span><span class="token function">type</span>=<span class="token string">'json'</span> <span class="token operator">-</span>p <span class="token string">'[{"op": "replace", "path": "/spec/predictor/canaryTrafficPercent", "value": 0}]'</span>
</code></pre>
<pre><code class="prism language-powershell">curl <span class="token operator">-</span>v <span class="token operator">-</span>H <span class="token string">"Host: latest-torchserve-predictor-default.default.example.com"</span> http:<span class="token operator">/</span><span class="token operator">/</span>$<span class="token punctuation">{</span>INGRESS_HOST<span class="token punctuation">}</span>:$<span class="token punctuation">{</span>INGRESS_PORT<span class="token punctuation">}</span><span class="token operator">/</span>v1/models/$<span class="token punctuation">{</span>MODEL_NAME<span class="token punctuation">}</span>:predict <span class="token operator">-</span>d @<span class="token punctuation">.</span><span class="token operator">/</span>mnist<span class="token punctuation">.</span>json
</code></pre>
<p><code>期望输出</code></p>
<pre><code>{"id": "d3b15cad-50a2-4eaf-80ce-8b0a428bd298", "model_name": "mnist", "model_version": "1.0", "outputs": [{"name": "predict", "shape": [1], "datatype": "INT64", "data": [1]}]}
</code></pre>
<p>在测试和验证了新模型后，您现在可以将canaryTrafficPercent提高到100，以将流量完全推送到新版本，现在latestRolledoutRevision变为torchserver-predictor-default-00002，而previousRolledoutRevision变成torchserver-predictor-default-00001。<br> <code>kubectl</code></p>
<pre><code class="prism language-powershell">kubectl patch isvc torchserve <span class="token operator">--</span><span class="token function">type</span>=<span class="token string">'json'</span> <span class="token operator">-</span>p <span class="token string">'[{"op": "replace", "path": "/spec/predictor/canaryTrafficPercent", "value": 100}]'</span>
</code></pre>
<p>检查流量状态：</p>
<pre><code class="prism language-powershell">kubectl get isvc torchserve <span class="token operator">-</span>ojsonpath=<span class="token string">'{.status.components}'</span>
</code></pre>
<p><code>期望输出</code></p>
<pre><code>{
"predictor": {
"address": {
"url": "http://torchserve-predictor-default.default.svc.cluster.local"
},
"latestCreatedRevision": "torchserve-predictor-default-00002",
"latestReadyRevision": "torchserve-predictor-default-00002",
"latestRolledoutRevision": "torchserve-predictor-default-00002",
"previousRolledoutRevision": "torchserve-predictor-default-00001",
"traffic": [
{
"latestRevision": true,
"percent": 100,
"revisionName": "torchserve-predictor-default-00002",
"tag": "latest",
"url": "http://latest-torchserve-predictor-default.default.example.com"
},
],
"url": "http://torchserve-predictor-default.default.example.com"
}
}
</code></pre>
<h4>回滚模型</h4>
<p>如果流量移动到新版本后，新模型版本不起作用，您仍然可以将canaryTrafficPercent修补为0，并将流量移回以前滚动的模型，即torchserver-predictor-default-00001。</p>
<p><code>kubectl</code></p>
<pre><code class="prism language-powershell">kubectl patch isvc torchserve <span class="token operator">--</span><span class="token function">type</span>=<span class="token string">'json'</span> <span class="token operator">-</span>p <span class="token string">'[{"op": "replace", "path": "/spec/predictor/canaryTrafficPercent", "value": 0}]'</span>
</code></pre>
<p>检查流量状态：</p>
<pre><code class="prism language-powershell">kubectl get isvc torchserve <span class="token operator">-</span>ojsonpath=<span class="token string">'{.status.components}'</span>
</code></pre>
<p><code>期望输出</code></p>
<pre><code>{
"predictor": {
"address": {
"url": "http://torchserve-predictor-default.default.svc.cluster.local"
},
"latestCreatedRevision": "torchserve-predictor-default-00002",
"latestReadyRevision": "torchserve-predictor-default-00002",
"latestRolledoutRevision": "torchserve-predictor-default-00001",
"previousRolledoutRevision": "torchserve-predictor-default-00001",
"traffic": [
{
"latestRevision": true,
"percent": 0,
"revisionName": "torchserve-predictor-default-00002",
"tag": "latest",
"url": "http://latest-torchserve-predictor-default.default.example.com"
},
{
"latestRevision": false,
"percent": 100,
"revisionName": "torchserve-predictor-default-00001",
"tag": "prev",
"url": "http://prev-torchserve-predictor-default.default.example.com"
}
],
"url": "http://torchserve-predictor-default.default.example.com"
}
}
</code></pre>
<h3>监控</h3>
<p>Metrics Exposure and Grafana Dashboard Setup</p>
<h2>使用推理服务部署Scikit学习模型</h2>
<p>本示例将引导您了解如何利用推理服务CRD的v1beta1版本部署scikit-learn模型。请注意，默认情况下，v1beta1版本将通过与现有V1 Dataplane兼容的API公开您的模型。但是，本示例将向您展示如何通过与新的V2 Dataplane兼容的API为模型提供服务。</p>
<h3>训练</h3>
<p>第一步是训练一个样本scikit学习模型。请注意，该模型将保存为model.joblib。</p>
<pre><code class="prism language-py"><span class="token keyword">from</span> sklearn <span class="token keyword">import</span> svm
<span class="token keyword">from</span> sklearn <span class="token keyword">import</span> datasets
<span class="token keyword">from</span> joblib <span class="token keyword">import</span> dump
iris <span class="token operator">=</span> datasets<span class="token punctuation">.</span>load_iris<span class="token punctuation">(</span><span class="token punctuation">)</span>
X<span class="token punctuation">,</span> y <span class="token operator">=</span> iris<span class="token punctuation">.</span>data<span class="token punctuation">,</span> iris<span class="token punctuation">.</span>target
clf <span class="token operator">=</span> svm<span class="token punctuation">.</span>SVC<span class="token punctuation">(</span>gamma<span class="token operator">=</span><span class="token string">'scale'</span><span class="token punctuation">)</span>
clf<span class="token punctuation">.</span>fit<span class="token punctuation">(</span>X<span class="token punctuation">,</span> y<span class="token punctuation">)</span>
dump<span class="token punctuation">(</span>clf<span class="token punctuation">,</span> <span class="token string">'model.joblib'</span><span class="token punctuation">)</span>
</code></pre>
<h3>本地测试</h3>
<p>一旦您获得了模型序列化的model.joblib，我们就可以使用MLServer来启动本地服务器。有关MLServer的更多详细信息，请随时查看SKLearn示例文档。</p>
<blockquote>
<p>笔记<br> 此步骤是可选的，仅用于测试，可以直接使用InfenceService进行部署。</p>
</blockquote>
<h4>先决条件</h4>
<p>首先，要在本地使用MLServer，首先需要在本地环境以及SKLearn运行时中安装MLServer包。</p>
<pre><code>pip install mlserver mlserver-sklearn
</code></pre>
<h4>模型设置</h4>
<p>下一步将提供一些模型设置，以便MLServer知道：</p>
<ul>
<li>为您的模型提供服务的推理运行时（即mlserver_sklearn.SKLearnModel）</li>
<li>型号的名称和版本</li>
</ul>
<p>这些可以通过环境变量或创建本地model-settings.json文件来指定：</p>
<pre><code class="prism language-python"><span class="token punctuation">{</span>
<span class="token string">"name"</span><span class="token punctuation">:</span> <span class="token string">"sklearn-iris"</span><span class="token punctuation">,</span>
<span class="token string">"version"</span><span class="token punctuation">:</span> <span class="token string">"v1.0.0"</span><span class="token punctuation">,</span>
<span class="token string">"implementation"</span><span class="token punctuation">:</span> <span class="token string">"mlserver_sklearn.SKLearnModel"</span>
<span class="token punctuation">}</span>
</code></pre>
<p>请注意，当您部署模型时，KServe已经注入了一些合理的默认值，这样它就可以开箱即用，而无需任何进一步的配置。但是，您仍然可以通过提供类似于本地文件的model-settings.json文件来覆盖这些默认值。您甚至可以提供一组model-settings.json文件来加载多个模型。</p>
<h4>本地服务模型</h4>
<p>有了本地安装的mlserver软件包和本地model-settings.json文件，您现在应该可以启动我们的服务器了：</p>
<pre><code>mlserver start .
</code></pre>
<h3>使用推理服务进行部署</h3>
<p>最后，您将使用KServe来部署经过训练的模型。为此，您只需要使用推理服务CRD的v1beta1版本，并将protocolVersion字段设置为v2。<br> <code>旧框架</code></p>
<pre><code class="prism language-yaml"><span class="token key atrule">apiVersion</span><span class="token punctuation">:</span> <span class="token string">"serving.kserve.io/v1beta1"</span>
<span class="token key atrule">kind</span><span class="token punctuation">:</span> <span class="token string">"InferenceService"</span>
<span class="token key atrule">metadata</span><span class="token punctuation">:</span>
<span class="token key atrule">name</span><span class="token punctuation">:</span> <span class="token string">"sklearn-irisv2"</span>
<span class="token key atrule">spec</span><span class="token punctuation">:</span>
<span class="token key atrule">predictor</span><span class="token punctuation">:</span>
<span class="token key atrule">sklearn</span><span class="token punctuation">:</span>
<span class="token key atrule">protocolVersion</span><span class="token punctuation">:</span> <span class="token string">"v2"</span>
<span class="token key atrule">storageUri</span><span class="token punctuation">:</span> <span class="token string">"gs://seldon-models/sklearn/mms/lr_model"</span>
</code></pre>
<p><code>新框架</code></p>
<pre><code class="prism language-yaml"><span class="token key atrule">apiVersion</span><span class="token punctuation">:</span> <span class="token string">"serving.kserve.io/v1beta1"</span>
<span class="token key atrule">kind</span><span class="token punctuation">:</span> <span class="token string">"InferenceService"</span>
<span class="token key atrule">metadata</span><span class="token punctuation">:</span>
<span class="token key atrule">name</span><span class="token punctuation">:</span> <span class="token string">"sklearn-irisv2"</span>
<span class="token key atrule">spec</span><span class="token punctuation">:</span>
<span class="token key atrule">predictor</span><span class="token punctuation">:</span>
<span class="token key atrule">model</span><span class="token punctuation">:</span>
<span class="token key atrule">modelFormat</span><span class="token punctuation">:</span>
<span class="token key atrule">name</span><span class="token punctuation">:</span> sklearn
<span class="token key atrule">runtime</span><span class="token punctuation">:</span> kserve<span class="token punctuation">-</span>mlserver
<span class="token key atrule">storageUri</span><span class="token punctuation">:</span> <span class="token string">"gs://seldon-models/sklearn/mms/lr_model"</span>
</code></pre>
<p>请注意，这做出了以下假设：</p>
<ul>
<li>您的模型权重（即model.joblib文件）已经上传到“模型存储库”（本例中为GCS），并且可以作为gs://seldon models/skrearn/iris访问。</li>
<li>有一个K8s集群可用，可以通过kubectl访问。</li>
<li>KServe已安装在您的群集中。</li>
</ul>
<p><code>kubectl</code></p>
<pre><code class="prism language-powershell">kubectl apply <span class="token operator">-</span>f <span class="token punctuation">.</span><span class="token operator">/</span>sklearn<span class="token punctuation">.</span>yaml
</code></pre>
<h3>测试部署的模型</h3>
<p>现在，您可以通过发送示例请求来测试已部署的模型。<br> 请注意，此请求需要遵循V2数据平面协议。您可以在下面看到一个负载示例：</p>
<pre><code class="prism language-python"><span class="token punctuation">{</span>
<span class="token string">"inputs"</span><span class="token punctuation">:</span> <span class="token punctuation">[</span>
<span class="token punctuation">{</span>
<span class="token string">"name"</span><span class="token punctuation">:</span> <span class="token string">"input-0"</span><span class="token punctuation">,</span>
<span class="token string">"shape"</span><span class="token punctuation">:</span> <span class="token punctuation">[</span><span class="token number">2</span><span class="token punctuation">,</span> <span class="token number">4</span><span class="token punctuation">]</span><span class="token punctuation">,</span>
<span class="token string">"datatype"</span><span class="token punctuation">:</span> <span class="token string">"FP32"</span><span class="token punctuation">,</span>
<span class="token string">"data"</span><span class="token punctuation">:</span> <span class="token punctuation">[</span>
<span class="token punctuation">[</span><span class="token number">6.8</span><span class="token punctuation">,</span> <span class="token number">2.8</span><span class="token punctuation">,</span> <span class="token number">4.8</span><span class="token punctuation">,</span> <span class="token number">1.4</span><span class="token punctuation">]</span><span class="token punctuation">,</span>
<span class="token punctuation">[</span><span class="token number">6.0</span><span class="token punctuation">,</span> <span class="token number">3.4</span><span class="token punctuation">,</span> <span class="token number">4.5</span><span class="token punctuation">,</span> <span class="token number">1.6</span><span class="token punctuation">]</span>
<span class="token punctuation">]</span>
<span class="token punctuation">}</span>
<span class="token punctuation">]</span>
<span class="token punctuation">}</span>
</code></pre>
<p>现在，假设您的入口可以在${INGRESS_HOST}: ${INGRESS_PORT}访问，或者您可以按照此说明查找入口IP和端口。</p>
<p>您可以使用curl将推理请求发送为：</p>
<pre><code class="prism language-powershell">SERVICE_HOSTNAME=$<span class="token punctuation">(</span>kubectl get inferenceservice sklearn-irisv2 <span class="token operator">-</span>o jsonpath=<span class="token string">'{.status.url}'</span> <span class="token punctuation">|</span> cut <span class="token operator">-</span>d <span class="token string">"/"</span> <span class="token operator">-</span>f 3<span class="token punctuation">)</span>
curl <span class="token operator">-</span>v \
<span class="token operator">-</span>H <span class="token string">"Host: ${SERVICE_HOSTNAME}"</span> \
<span class="token operator">-</span>H <span class="token string">"Content-Type: application/json"</span> \
<span class="token operator">-</span>d @<span class="token punctuation">.</span><span class="token operator">/</span>iris-input<span class="token punctuation">.</span>json \
http:<span class="token operator">/</span><span class="token operator">/</span>$<span class="token punctuation">{</span>INGRESS_HOST<span class="token punctuation">}</span>:$<span class="token punctuation">{</span>INGRESS_PORT<span class="token punctuation">}</span><span class="token operator">/</span>v2/models/sklearn-irisv2/infer
</code></pre>
<p><code>期望输出</code></p>
<pre><code class="prism language-py"><span class="token punctuation">{</span>
<span class="token string">"id"</span><span class="token punctuation">:</span> <span class="token string">"823248cc-d770-4a51-9606-16803395569c"</span><span class="token punctuation">,</span>
<span class="token string">"model_name"</span><span class="token punctuation">:</span> <span class="token string">"sklearn-irisv2"</span><span class="token punctuation">,</span>
<span class="token string">"model_version"</span><span class="token punctuation">:</span> <span class="token string">"v1.0.0"</span><span class="token punctuation">,</span>
<span class="token string">"outputs"</span><span class="token punctuation">:</span> <span class="token punctuation">[</span>
<span class="token punctuation">{</span>
<span class="token string">"data"</span><span class="token punctuation">:</span> <span class="token punctuation">[</span><span class="token number">1</span><span class="token punctuation">,</span> <span class="token number">1</span><span class="token punctuation">]</span><span class="token punctuation">,</span>
<span class="token string">"datatype"</span><span class="token punctuation">:</span> <span class="token string">"INT64"</span><span class="token punctuation">,</span>
<span class="token string">"name"</span><span class="token punctuation">:</span> <span class="token string">"predict"</span><span class="token punctuation">,</span>
<span class="token string">"parameters"</span><span class="token punctuation">:</span> null<span class="token punctuation">,</span>
<span class="token string">"shape"</span><span class="token punctuation">:</span> <span class="token punctuation">[</span><span class="token number">2</span><span class="token punctuation">]</span>
<span class="token punctuation">}</span>
<span class="token punctuation">]</span>
<span class="token punctuation">}</span>
</code></pre>
<h2>使用推理服务部署XGBoost模型</h2>
<p>本示例将引导您了解如何利用推理服务CRD的v1beta1版本部署xgboost模型。请注意，默认情况下，v1beta1版本将通过与现有V1Dataplane兼容的API公开您的模型。但是，本示例将向您展示如何通过与新的V2数据平面兼容的API为模型提供服务。</p>
<h3>训练</h3>
<p>第一步是训练一个样例xgboost模型。我们将此模型另存为model.bst。</p>
<pre><code class="prism language-python"><span class="token keyword">import</span> xgboost <span class="token keyword">as</span> xgb
<span class="token keyword">from</span> sklearn<span class="token punctuation">.</span>datasets <span class="token keyword">import</span> load_iris
<span class="token keyword">import</span> os
model_dir <span class="token operator">=</span> <span class="token string">"."</span>
BST_FILE <span class="token operator">=</span> <span class="token string">"model.bst"</span>
iris <span class="token operator">=</span> load_iris<span class="token punctuation">(</span><span class="token punctuation">)</span>
y <span class="token operator">=</span> iris<span class="token punctuation">[</span><span class="token string">'target'</span><span class="token punctuation">]</span>
X <span class="token operator">=</span> iris<span class="token punctuation">[</span><span class="token string">'data'</span><span class="token punctuation">]</span>
dtrain <span class="token operator">=</span> xgb<span class="token punctuation">.</span>DMatrix<span class="token punctuation">(</span>X<span class="token punctuation">,</span> label<span class="token operator">=</span>y<span class="token punctuation">)</span>
param <span class="token operator">=</span> <span class="token punctuation">{</span><span class="token string">'max_depth'</span><span class="token punctuation">:</span> <span class="token number">6</span><span class="token punctuation">,</span>
<span class="token string">'eta'</span><span class="token punctuation">:</span> <span class="token number">0.1</span><span class="token punctuation">,</span>
<span class="token string">'silent'</span><span class="token punctuation">:</span> <span class="token number">1</span><span class="token punctuation">,</span>
<span class="token string">'nthread'</span><span class="token punctuation">:</span> <span class="token number">4</span><span class="token punctuation">,</span>
<span class="token string">'num_class'</span><span class="token punctuation">:</span> <span class="token number">10</span><span class="token punctuation">,</span>
<span class="token string">'objective'</span><span class="token punctuation">:</span> <span class="token string">'multi:softmax'</span>
<span class="token punctuation">}</span>
xgb_model <span class="token operator">=</span> xgb<span class="token punctuation">.</span>train<span class="token punctuation">(</span>params<span class="token operator">=</span>param<span class="token punctuation">,</span> dtrain<span class="token operator">=</span>dtrain<span class="token punctuation">)</span>
model_file <span class="token operator">=</span> os<span class="token punctuation">.</span>path<span class="token punctuation">.</span>join<span class="token punctuation">(</span><span class="token punctuation">(</span>model_dir<span class="token punctuation">)</span><span class="token punctuation">,</span> BST_FILE<span class="token punctuation">)</span>
xgb_model<span class="token punctuation">.</span>save_model<span class="token punctuation">(</span>model_file<span class="token punctuation">)</span>
</code></pre>
<h3>本地测试</h3>
<p>一旦我们的model.bst模型串行化，我们就可以使用MLServer来启动本地服务器。有关MLServer的更多详细信息，请随时查看他们文档中的XGBoost示例。</p>
<blockquote>
<p>请注意，此步骤是可选的，仅用于测试。您可以直接部署经过训练的模型。</p>
</blockquote>
<h4>先决条件</h4>
<p>首先，要在本地使用MLServer，首先需要在本地环境以及XGBoost运行时中安装MLServer包。</p>
<pre><code>pip install mlserver mlserver-xgboost
</code></pre>
<h4>模型设置</h4>
<p>下一步将提供一些模型设置，以便MLServer知道：</p>
<ul>
<li>我们希望模型使用的推理运行时（即mlserver_xgboost.XGBoostModel）</li>
<li>我们型号的名称和版本</li>
</ul>
<p>这些可以通过环境变量或创建本地model-settings.json文件来指定：</p>
<pre><code class="prism language-python"><span class="token punctuation">{</span>
<span class="token string">"name"</span><span class="token punctuation">:</span> <span class="token string">"xgboost-iris"</span><span class="token punctuation">,</span>
<span class="token string">"version"</span><span class="token punctuation">:</span> <span class="token string">"v1.0.0"</span><span class="token punctuation">,</span>
<span class="token string">"implementation"</span><span class="token punctuation">:</span> <span class="token string">"mlserver_xgboost.XGBoostModel"</span>
<span class="token punctuation">}</span>
</code></pre>
<p>请注意，当我们部署我们的模型时，KServe已经注入了一些合理的默认值，这样它就可以在没有任何进一步配置的情况下开箱即用。但是，您仍然可以通过提供类似于本地文件的model-settings.json文件来覆盖这些默认值。您甚至可以提供一组model-settings.json文件来加载多个模型。</p>
<h4>在本地为我们的模型服务</h4>
<p>有了本地安装的mlserver包和本地model-settings.json文件，我们现在应该可以启动服务器了：</p>
<pre><code>mlserver start .
</code></pre>
<h3>使用推理服务进行部署</h3>
<p>最后，我们将使用KServe来部署我们经过训练的模型。为此，我们只需要使用推理服务CRD的v1beta1版本，并将protocolVersion字段设置为v2。<br> <code>旧框架</code></p>
<pre><code class="prism language-yaml"><span class="token key atrule">apiVersion</span><span class="token punctuation">:</span> <span class="token string">"serving.kserve.io/v1beta1"</span>
<span class="token key atrule">kind</span><span class="token punctuation">:</span> <span class="token string">"InferenceService"</span>
<span class="token key atrule">metadata</span><span class="token punctuation">:</span>
<span class="token key atrule">name</span><span class="token punctuation">:</span> <span class="token string">"xgboost-iris"</span>
<span class="token key atrule">spec</span><span class="token punctuation">:</span>
<span class="token key atrule">predictor</span><span class="token punctuation">:</span>
<span class="token key atrule">xgboost</span><span class="token punctuation">:</span>
<span class="token key atrule">protocolVersion</span><span class="token punctuation">:</span> <span class="token string">"v2"</span>
<span class="token key atrule">storageUri</span><span class="token punctuation">:</span> <span class="token string">"gs://kfserving-examples/models/xgboost/iris"</span>
</code></pre>
<p><code>新框架</code></p>
<pre><code class="prism language-yaml"><span class="token key atrule">apiVersion</span><span class="token punctuation">:</span> <span class="token string">"serving.kserve.io/v1beta1"</span>
<span class="token key atrule">kind</span><span class="token punctuation">:</span> <span class="token string">"InferenceService"</span>
<span class="token key atrule">metadata</span><span class="token punctuation">:</span>
<span class="token key atrule">name</span><span class="token punctuation">:</span> <span class="token string">"xgboost-iris"</span>
<span class="token key atrule">spec</span><span class="token punctuation">:</span>
<span class="token key atrule">predictor</span><span class="token punctuation">:</span>
<span class="token key atrule">model</span><span class="token punctuation">:</span>
<span class="token key atrule">modelFormat</span><span class="token punctuation">:</span>
<span class="token key atrule">name</span><span class="token punctuation">:</span> xgboost
<span class="token key atrule">runtime</span><span class="token punctuation">:</span> kserve<span class="token punctuation">-</span>mlserver
<span class="token key atrule">storageUri</span><span class="token punctuation">:</span> <span class="token string">"gs://kfserving-examples/models/xgboost/iris"</span>
</code></pre>
<p>请注意，这做出了以下假设：</p>
<ul>
<li>您的模型权重（即您的model.bst文件）已上传到“模型存储库”（本例中为GCS），<br> 可以作为gs://k服务示例/models/xgboost/iris访问。</li>
<li>有一个K8s集群可用，可以通过kubectl访问。</li>
<li>KServe已安装在您的群集中。</li>
</ul>
<p>假设我们已经安装了KServe，可以通过kubectl访问集群，我们可以将我们的模型部署为：</p>
<pre><code>kubectl apply -f xgboost.yaml
</code></pre>
<h3>测试部署的模型</h3>
<p>现在，我们可以通过发送一个示例请求来测试我们部署的模型。<br> 请注意，此请求需要遵循V2数据平面协议。您可以在下面看到一个负载示例：</p>
<pre><code class="prism language-python"><span class="token punctuation">{</span>
<span class="token string">"inputs"</span><span class="token punctuation">:</span> <span class="token punctuation">[</span>
<span class="token punctuation">{</span>
<span class="token string">"name"</span><span class="token punctuation">:</span> <span class="token string">"input-0"</span><span class="token punctuation">,</span>
<span class="token string">"shape"</span><span class="token punctuation">:</span> <span class="token punctuation">[</span><span class="token number">2</span><span class="token punctuation">,</span> <span class="token number">4</span><span class="token punctuation">]</span><span class="token punctuation">,</span>
<span class="token string">"datatype"</span><span class="token punctuation">:</span> <span class="token string">"FP32"</span><span class="token punctuation">,</span>
<span class="token string">"data"</span><span class="token punctuation">:</span> <span class="token punctuation">[</span>
<span class="token punctuation">[</span><span class="token number">6.8</span><span class="token punctuation">,</span> <span class="token number">2.8</span><span class="token punctuation">,</span> <span class="token number">4.8</span><span class="token punctuation">,</span> <span class="token number">1.4</span><span class="token punctuation">]</span><span class="token punctuation">,</span>
<span class="token punctuation">[</span><span class="token number">6.0</span><span class="token punctuation">,</span> <span class="token number">3.4</span><span class="token punctuation">,</span> <span class="token number">4.5</span><span class="token punctuation">,</span> <span class="token number">1.6</span><span class="token punctuation">]</span>
<span class="token punctuation">]</span>
<span class="token punctuation">}</span>
<span class="token punctuation">]</span>
<span class="token punctuation">}</span>
</code></pre>
<p>现在，假设我们的入口可以在${INGRESS_HOST}: ${INGRESS_PORT}访问，我们可以使用curl将我们的推理请求发送为：</p>
<blockquote>
<p>您可以按照以下说明查找入口IP和端口。</p>
</blockquote>
<pre><code class="prism language-powershell">SERVICE_HOSTNAME=$<span class="token punctuation">(</span>kubectl get inferenceservice xgboost-iris <span class="token operator">-</span>o jsonpath=<span class="token string">'{.status.url}'</span> <span class="token punctuation">|</span> cut <span class="token operator">-</span>d <span class="token string">"/"</span> <span class="token operator">-</span>f 3<span class="token punctuation">)</span>
curl <span class="token operator">-</span>v \
<span class="token operator">-</span>H <span class="token string">"Host: ${SERVICE_HOSTNAME}"</span> \
<span class="token operator">-</span>H <span class="token string">"Content-Type: application/json"</span> \
<span class="token operator">-</span>d @<span class="token punctuation">.</span><span class="token operator">/</span>iris-input<span class="token punctuation">.</span>json \
http:<span class="token operator">/</span><span class="token operator">/</span>$<span class="token punctuation">{</span>INGRESS_HOST<span class="token punctuation">}</span>:$<span class="token punctuation">{</span>INGRESS_PORT<span class="token punctuation">}</span><span class="token operator">/</span>v2/models/xgboost-iris/infer
</code></pre>
<p>输出将类似于：<br> <code>期望输出</code></p>
<pre><code class="prism language-python"><span class="token punctuation">{</span>
<span class="token string">"id"</span><span class="token punctuation">:</span> <span class="token string">"4e546709-0887-490a-abd6-00cbc4c26cf4"</span><span class="token punctuation">,</span>
<span class="token string">"model_name"</span><span class="token punctuation">:</span> <span class="token string">"xgboost-iris"</span><span class="token punctuation">,</span>
<span class="token string">"model_version"</span><span class="token punctuation">:</span> <span class="token string">"v1.0.0"</span><span class="token punctuation">,</span>
<span class="token string">"outputs"</span><span class="token punctuation">:</span> <span class="token punctuation">[</span>
<span class="token punctuation">{</span>
<span class="token string">"data"</span><span class="token punctuation">:</span> <span class="token punctuation">[</span><span class="token number">1.0</span><span class="token punctuation">,</span> <span class="token number">1.0</span><span class="token punctuation">]</span><span class="token punctuation">,</span>
<span class="token string">"datatype"</span><span class="token punctuation">:</span> <span class="token string">"FP32"</span><span class="token punctuation">,</span>
<span class="token string">"name"</span><span class="token punctuation">:</span> <span class="token string">"predict"</span><span class="token punctuation">,</span>
<span class="token string">"parameters"</span><span class="token punctuation">:</span> null<span class="token punctuation">,</span>
<span class="token string">"shape"</span><span class="token punctuation">:</span> <span class="token punctuation">[</span><span class="token number">2</span><span class="token punctuation">]</span>
<span class="token punctuation">}</span>
<span class="token punctuation">]</span>
<span class="token punctuation">}</span>
</code></pre>
<h2>使用推理服务部署PMML模型</h2>
<p>PMML，或称预测模型标记语言，是一种用于描述数据挖掘和统计模型的XML格式，包括模型的输入、用于为数据挖掘准备数据的转换，以及定义模型本身的参数。在这个例子中，我们展示了如何在推理服务上提供PMML格式模型。</p>
<h3>创建推理服务</h3>
<p><code>旧框架</code></p>
<pre><code class="prism language-yaml"><span class="token key atrule">apiVersion</span><span class="token punctuation">:</span> <span class="token string">"serving.kserve.io/v1beta1"</span>
<span class="token key atrule">kind</span><span class="token punctuation">:</span> <span class="token string">"InferenceService"</span>
<span class="token key atrule">metadata</span><span class="token punctuation">:</span>
<span class="token key atrule">name</span><span class="token punctuation">:</span> <span class="token string">"pmml-demo"</span>
<span class="token key atrule">spec</span><span class="token punctuation">:</span>
<span class="token key atrule">predictor</span><span class="token punctuation">:</span>
<span class="token key atrule">pmml</span><span class="token punctuation">:</span>
<span class="token key atrule">storageUri</span><span class="token punctuation">:</span> gs<span class="token punctuation">:</span>//kfserving<span class="token punctuation">-</span>examples/models/pmml
</code></pre>
<p><code>新框架</code></p>
<pre><code class="prism language-yaml"><span class="token key atrule">apiVersion</span><span class="token punctuation">:</span> <span class="token string">"serving.kserve.io/v1beta1"</span>
<span class="token key atrule">kind</span><span class="token punctuation">:</span> <span class="token string">"InferenceService"</span>
<span class="token key atrule">metadata</span><span class="token punctuation">:</span>
<span class="token key atrule">name</span><span class="token punctuation">:</span> <span class="token string">"pmml-demo"</span>
<span class="token key atrule">spec</span><span class="token punctuation">:</span>
<span class="token key atrule">predictor</span><span class="token punctuation">:</span>
<span class="token key atrule">model</span><span class="token punctuation">:</span>
<span class="token key atrule">modelFormat</span><span class="token punctuation">:</span>
<span class="token key atrule">name</span><span class="token punctuation">:</span> pmml
<span class="token key atrule">storageUri</span><span class="token punctuation">:</span> <span class="token string">"gs://kfserving-examples/models/pmml"</span>
</code></pre>
<p>使用上述yaml创建推理服务</p>
<pre><code class="prism language-powershell">kubectl apply <span class="token operator">-</span>f pmml<span class="token punctuation">.</span>yaml
</code></pre>
<p><code>期望输出</code></p>
<pre><code>$ inferenceservice.serving.kserve.io/pmml-demo created
</code></pre>
<blockquote>
<p>警告<br> pmmlserver基于Py4J，不支持多进程模式，因此我们无法设置spec.productor.contenerConcurrency。如果您想扩展pmmlserver以提高预测性能，您应该将推理服务的resources.limits.cu设置为1，并扩展副本大小。</p>
</blockquote>
<h3>运行预测</h3>
<p>第一步是确定入口IP和端口，并设置INGRESS_HOST和INGRESS_PORT</p>
<pre><code class="prism language-powershell">MODEL_NAME=pmml-demo
INPUT_PATH=@<span class="token punctuation">.</span><span class="token operator">/</span>pmml-input<span class="token punctuation">.</span>json
SERVICE_HOSTNAME=$<span class="token punctuation">(</span>kubectl get inferenceservice pmml-demo <span class="token operator">-</span>o jsonpath=<span class="token string">'{.status.url}'</span> <span class="token punctuation">|</span> cut <span class="token operator">-</span>d <span class="token string">"/"</span> <span class="token operator">-</span>f 3<span class="token punctuation">)</span>
curl <span class="token operator">-</span>v <span class="token operator">-</span>H <span class="token string">"Host: ${SERVICE_HOSTNAME}"</span> http:<span class="token operator">/</span><span class="token operator">/</span>$<span class="token punctuation">{</span>INGRESS_HOST<span class="token punctuation">}</span>:$<span class="token punctuation">{</span>INGRESS_PORT<span class="token punctuation">}</span><span class="token operator">/</span>v1/models/<span class="token variable">$MODEL_NAME</span>:predict <span class="token operator">-</span>d <span class="token variable">$INPUT_PATH</span>
</code></pre>
<p><code>期望输出</code></p>
<pre><code class="prism language-powershell"><span class="token operator">*</span> TCP_NODELAY <span class="token function">set</span>
<span class="token operator">*</span> Connected to localhost <span class="token punctuation">(</span>::1<span class="token punctuation">)</span> port 8081 <span class="token punctuation">(</span><span class="token comment">#0)</span>
> POST <span class="token operator">/</span>v1/models/pmml-demo:predict HTTP/1<span class="token punctuation">.</span>1
> Host: pmml-demo<span class="token punctuation">.</span>default<span class="token punctuation">.</span>example<span class="token punctuation">.</span>com
> User-Agent: curl/7<span class="token punctuation">.</span>64<span class="token punctuation">.</span>1
> Accept: <span class="token operator">*</span><span class="token operator">/</span><span class="token operator">*</span>
> Content-Length: 45
> Content-<span class="token function">Type</span>: application/x-www-form-urlencoded
>
<span class="token operator">*</span> upload completely sent off: 45 out of 45 bytes
< HTTP/1<span class="token punctuation">.</span>1 200 OK
< content-length: 39
< content-<span class="token function">type</span>: application/json<span class="token punctuation">;</span> charset=UTF-8
< date: Sun<span class="token punctuation">,</span> 18 Oct 2020 15:50:02 GMT
< server: istio-envoy
< x-envoy-upstream-service-time: 12
<
<span class="token operator">*</span> Connection <span class="token comment">#0 to host localhost left intact</span>
<span class="token punctuation">{</span><span class="token string">"predictions"</span>: <span class="token punctuation">[</span><span class="token punctuation">{</span><span class="token string">'Species'</span>: <span class="token string">'setosa'</span><span class="token punctuation">,</span> <span class="token string">'Probability_setosa'</span>: 1<span class="token punctuation">.</span>0<span class="token punctuation">,</span> <span class="token string">'Probability_versicolor'</span>: 0<span class="token punctuation">.</span>0<span class="token punctuation">,</span> <span class="token string">'Probability_virginica'</span>: 0<span class="token punctuation">.</span>0<span class="token punctuation">,</span> <span class="token string">'Node_Id'</span>: <span class="token string">'2'</span><span class="token punctuation">}</span><span class="token punctuation">]</span><span class="token punctuation">}</span><span class="token operator">*</span> Closing connection 0
</code></pre>
<h2>使用PMML推理服务部署Spark MLlib模型</h2>
<h3>设置</h3>
<p>1.安装pyspark 3.0.x和pyspark2pmml</p>
<pre><code>pip install pyspark~=3.0.0
pip install pyspark2pmml
</code></pre>
<p>2.获取JPMML-SparkML.jar</p>
<h3>训练Spark MLlib模型并导出到PMML文件</h3>
<p>使用–jars启动pyspark以指定JPMML-SparkML uber-JAR的位置</p>
<pre><code>pyspark --jars ./jpmml-sparkml-executable-1.6.3.jar
</code></pre>
<p>安装Spark ML管道</p>
<pre><code class="prism language-python"><span class="token keyword">from</span> pyspark<span class="token punctuation">.</span>ml <span class="token keyword">import</span> Pipeline
<span class="token keyword">from</span> pyspark<span class="token punctuation">.</span>ml<span class="token punctuation">.</span>classification <span class="token keyword">import</span> DecisionTreeClassifier
<span class="token keyword">from</span> pyspark<span class="token punctuation">.</span>ml<span class="token punctuation">.</span>feature <span class="token keyword">import</span> RFormula
df <span class="token operator">=</span> spark<span class="token punctuation">.</span>read<span class="token punctuation">.</span>csv<span class="token punctuation">(</span><span class="token string">"Iris.csv"</span><span class="token punctuation">,</span> header <span class="token operator">=</span> <span class="token boolean">True</span><span class="token punctuation">,</span> inferSchema <span class="token operator">=</span> <span class="token boolean">True</span><span class="token punctuation">)</span>
formula <span class="token operator">=</span> RFormula<span class="token punctuation">(</span>formula <span class="token operator">=</span> <span class="token string">"Species ~ ."</span><span class="token punctuation">)</span>
classifier <span class="token operator">=</span> DecisionTreeClassifier<span class="token punctuation">(</span><span class="token punctuation">)</span>
pipeline <span class="token operator">=</span> Pipeline<span class="token punctuation">(</span>stages <span class="token operator">=</span> <span class="token punctuation">[</span>formula<span class="token punctuation">,</span> classifier<span class="token punctuation">]</span><span class="token punctuation">)</span>
pipelineModel <span class="token operator">=</span> pipeline<span class="token punctuation">.</span>fit<span class="token punctuation">(</span>df<span class="token punctuation">)</span>
<span class="token keyword">from</span> pyspark2pmml <span class="token keyword">import</span> PMMLBuilder
pmmlBuilder <span class="token operator">=</span> PMMLBuilder<span class="token punctuation">(</span>sc<span class="token punctuation">,</span> df<span class="token punctuation">,</span> pipelineModel<span class="token punctuation">)</span>
pmmlBuilder<span class="token punctuation">.</span>buildFile<span class="token punctuation">(</span><span class="token string">"DecisionTreeIris.pmml"</span><span class="token punctuation">)</span>
</code></pre>
<p>将DecisionTreeRis.pmml上传到GCS存储桶，注意PMMLServer期望模型文件名为model.pmml</p>
<pre><code>gsutil cp ./DecisionTreeIris.pmml gs://$BUCKET_NAME/sparkpmml/model.pmml
</code></pre>
<h3>使用PMMLServer创建推理服务</h3>
<p>使用pmml预测器创建推理服务，并指定您上传到的存储桶位置的storageUri<br> <code>旧框架</code></p>
<pre><code class="prism language-yaml"><span class="token key atrule">apiVersion</span><span class="token punctuation">:</span> <span class="token string">"serving.kserve.io/v1beta1"</span>
<span class="token key atrule">kind</span><span class="token punctuation">:</span> <span class="token string">"InferenceService"</span>
<span class="token key atrule">metadata</span><span class="token punctuation">:</span>
<span class="token key atrule">name</span><span class="token punctuation">:</span> <span class="token string">"spark-pmml"</span>
<span class="token key atrule">spec</span><span class="token punctuation">:</span>
<span class="token key atrule">predictor</span><span class="token punctuation">:</span>
<span class="token key atrule">pmml</span><span class="token punctuation">:</span>
<span class="token key atrule">storageUri</span><span class="token punctuation">:</span> gs<span class="token punctuation">:</span>//kfserving<span class="token punctuation">-</span>examples/models/sparkpmml
</code></pre>
<p><code>新框架</code></p>
<pre><code class="prism language-yaml"><span class="token key atrule">apiVersion</span><span class="token punctuation">:</span> <span class="token string">"serving.kserve.io/v1beta1"</span>
<span class="token key atrule">kind</span><span class="token punctuation">:</span> <span class="token string">"InferenceService"</span>
<span class="token key atrule">metadata</span><span class="token punctuation">:</span>
<span class="token key atrule">name</span><span class="token punctuation">:</span> <span class="token string">"spark-pmml"</span>
<span class="token key atrule">spec</span><span class="token punctuation">:</span>
<span class="token key atrule">predictor</span><span class="token punctuation">:</span>
<span class="token key atrule">model</span><span class="token punctuation">:</span>
<span class="token key atrule">modelFormat</span><span class="token punctuation">:</span>
<span class="token key atrule">name</span><span class="token punctuation">:</span> pmml
<span class="token key atrule">storageUri</span><span class="token punctuation">:</span> gs<span class="token punctuation">:</span>//kfserving<span class="token punctuation">-</span>examples/models/sparkpmml
</code></pre>
<p>应用推理服务自定义资源</p>
<pre><code class="prism language-powershell">kubectl apply <span class="token operator">-</span>f spark_pmml<span class="token punctuation">.</span>yaml
</code></pre>
<p><code>期望输出</code></p>
<pre><code>$ inferenceservice.serving.kserve.io/spark-pmml created
</code></pre>
<p>等待推理服务准备就绪</p>
<pre><code class="prism language-powershell">kubectl wait <span class="token operator">--</span><span class="token keyword">for</span>=condition=Ready inferenceservice spark-pmml
inferenceservice<span class="token punctuation">.</span>serving<span class="token punctuation">.</span>kserve<span class="token punctuation">.</span>io/spark-pmml condition met
</code></pre>
<h4>运行预测</h4>
<p>第一步是确定入口IP和端口，并设置INGRESS_HOST和INGRESS_PORT</p>
<pre><code class="prism language-powershell">MODEL_NAME=spark-pmml
INPUT_PATH=@<span class="token punctuation">.</span><span class="token operator">/</span>pmml-input<span class="token punctuation">.</span>json
SERVICE_HOSTNAME=$<span class="token punctuation">(</span>kubectl get inferenceservice spark-pmml <span class="token operator">-</span>o jsonpath=<span class="token string">'{.status.url}'</span> <span class="token punctuation">|</span> cut <span class="token operator">-</span>d <span class="token string">"/"</span> <span class="token operator">-</span>f 3<span class="token punctuation">)</span>
curl <span class="token operator">-</span>v <span class="token operator">-</span>H <span class="token string">"Host: ${SERVICE_HOSTNAME}"</span> http:<span class="token operator">/</span><span class="token operator">/</span>$<span class="token punctuation">{</span>INGRESS_HOST<span class="token punctuation">}</span>:$<span class="token punctuation">{</span>INGRESS_PORT<span class="token punctuation">}</span><span class="token operator">/</span>v1/models/<span class="token variable">$MODEL_NAME</span>:predict <span class="token operator">-</span>d <span class="token variable">$INPUT_PATH</span>
</code></pre>
<p><code>期望输出</code></p>
<pre><code class="prism language-powershell"><span class="token operator">*</span> Connected to spark-pmml<span class="token punctuation">.</span>default<span class="token punctuation">.</span>35<span class="token punctuation">.</span>237<span class="token punctuation">.</span>217<span class="token punctuation">.</span>209<span class="token punctuation">.</span>xip<span class="token punctuation">.</span>io <span class="token punctuation">(</span>35<span class="token punctuation">.</span>237<span class="token punctuation">.</span>217<span class="token punctuation">.</span>209<span class="token punctuation">)</span> port 80 <span class="token punctuation">(</span><span class="token comment">#0)</span>
> POST <span class="token operator">/</span>v1/models/spark-pmml:predict HTTP/1<span class="token punctuation">.</span>1
> Host: spark-pmml<span class="token punctuation">.</span>default<span class="token punctuation">.</span>35<span class="token punctuation">.</span>237<span class="token punctuation">.</span>217<span class="token punctuation">.</span>209<span class="token punctuation">.</span>xip<span class="token punctuation">.</span>io
> User-Agent: curl/7<span class="token punctuation">.</span>73<span class="token punctuation">.</span>0
> Accept: <span class="token operator">*</span><span class="token operator">/</span><span class="token operator">*</span>
> Content-Length: 45
> Content-<span class="token function">Type</span>: application/x-www-form-urlencoded
>
<span class="token operator">*</span> upload completely sent off: 45 out of 45 bytes
<span class="token operator">*</span> Mark bundle as not supporting multiuse
< HTTP/1<span class="token punctuation">.</span>1 200 OK
< content-length: 39
< content-<span class="token function">type</span>: application/json<span class="token punctuation">;</span> charset=UTF-8
< date: Sun<span class="token punctuation">,</span> 07 Mar 2021 19:32:50 GMT
< server: istio-envoy
< x-envoy-upstream-service-time: 14
<
<span class="token operator">*</span> Connection <span class="token comment">#0 to host spark-pmml.default.35.237.217.209.xip.io left intact</span>
<span class="token punctuation">{</span><span class="token string">"predictions"</span>: <span class="token punctuation">[</span><span class="token punctuation">[</span>1<span class="token punctuation">.</span>0<span class="token punctuation">,</span> 0<span class="token punctuation">.</span>0<span class="token punctuation">,</span> 1<span class="token punctuation">.</span>0<span class="token punctuation">,</span> 0<span class="token punctuation">.</span>0<span class="token punctuation">]</span><span class="token punctuation">]</span><span class="token punctuation">}</span>
</code></pre>
<h2>使用推理服务部署LightGBM模型</h2>
<h3>训练LightGBM模型</h3>
<p>要测试LightGBM服务器，首先需要使用以下python代码训练一个简单的LightGBM模型。</p>
<pre><code class="prism language-python"><span class="token keyword">import</span> lightgbm <span class="token keyword">as</span> lgb
<span class="token keyword">from</span> sklearn<span class="token punctuation">.</span>datasets <span class="token keyword">import</span> load_iris
<span class="token keyword">import</span> os
model_dir <span class="token operator">=</span> <span class="token string">"."</span>
BST_FILE <span class="token operator">=</span> <span class="token string">"model.bst"</span>
iris <span class="token operator">=</span> load_iris<span class="token punctuation">(</span><span class="token punctuation">)</span>
y <span class="token operator">=</span> iris<span class="token punctuation">[</span><span class="token string">'target'</span><span class="token punctuation">]</span>
X <span class="token operator">=</span> iris<span class="token punctuation">[</span><span class="token string">'data'</span><span class="token punctuation">]</span>
dtrain <span class="token operator">=</span> lgb<span class="token punctuation">.</span>Dataset<span class="token punctuation">(</span>X<span class="token punctuation">,</span> label<span class="token operator">=</span>y<span class="token punctuation">,</span> feature_names<span class="token operator">=</span>iris<span class="token punctuation">[</span><span class="token string">'feature_names'</span><span class="token punctuation">]</span><span class="token punctuation">)</span>
params <span class="token operator">=</span> <span class="token punctuation">{</span>
<span class="token string">'objective'</span><span class="token punctuation">:</span><span class="token string">'multiclass'</span><span class="token punctuation">,</span>
<span class="token string">'metric'</span><span class="token punctuation">:</span><span class="token string">'softmax'</span><span class="token punctuation">,</span>
<span class="token string">'num_class'</span><span class="token punctuation">:</span> <span class="token number">3</span>
<span class="token punctuation">}</span>
lgb_model <span class="token operator">=</span> lgb<span class="token punctuation">.</span>train<span class="token punctuation">(</span>params<span class="token operator">=</span>params<span class="token punctuation">,</span> train_set<span class="token operator">=</span>dtrain<span class="token punctuation">)</span>
model_file <span class="token operator">=</span> os<span class="token punctuation">.</span>path<span class="token punctuation">.</span>join<span class="token punctuation">(</span>model_dir<span class="token punctuation">,</span> BST_FILE<span class="token punctuation">)</span>
lgb_model<span class="token punctuation">.</span>save_model<span class="token punctuation">(</span>model_file<span class="token punctuation">)</span>
</code></pre>
<h3>使用V1协议部署LightGBM模型</h3>
<h4>在本地测试模型</h4>
<p>使用经过训练的模型在本地安装并运行LightGBM服务器，并测试预测。</p>
<pre><code class="prism language-python">python <span class="token operator">-</span>m lgbserver <span class="token operator">-</span><span class="token operator">-</span>model_dir <span class="token operator">/</span>path<span class="token operator">/</span>to<span class="token operator">/</span>model_dir <span class="token operator">-</span><span class="token operator">-</span>model_name lgb
</code></pre>
<p>LightGBM服务器在本地启动后，我们可以通过发送推理请求来测试模型。</p>
<pre><code class="prism language-python"><span class="token keyword">import</span> requests
request <span class="token operator">=</span> <span class="token punctuation">{</span><span class="token string">'sepal_width_(cm)'</span><span class="token punctuation">:</span> <span class="token punctuation">{</span><span class="token number">0</span><span class="token punctuation">:</span> <span class="token number">3.5</span><span class="token punctuation">}</span><span class="token punctuation">,</span> <span class="token string">'petal_length_(cm)'</span><span class="token punctuation">:</span> <span class="token punctuation">{</span><span class="token number">0</span><span class="token punctuation">:</span> <span class="token number">1.4</span><span class="token punctuation">}</span><span class="token punctuation">,</span> <span class="token string">'petal_width_(cm)'</span><span class="token punctuation">:</span> <span class="token punctuation">{</span><span class="token number">0</span><span class="token punctuation">:</span> <span class="token number">0.2</span><span class="token punctuation">}</span><span class="token punctuation">,</span><span class="token string">'sepal_length_(cm)'</span><span class="token punctuation">:</span> <span class="token punctuation">{</span><span class="token number">0</span><span class="token punctuation">:</span> <span class="token number">5.1</span><span class="token punctuation">}</span> <span class="token punctuation">}</span>
formData <span class="token operator">=</span> <span class="token punctuation">{</span>
<span class="token string">'inputs'</span><span class="token punctuation">:</span> <span class="token punctuation">[</span>request<span class="token punctuation">]</span>
<span class="token punctuation">}</span>
res <span class="token operator">=</span> requests<span class="token punctuation">.</span>post<span class="token punctuation">(</span><span class="token string">'http://localhost:8080/v1/models/lgb:predict'</span><span class="token punctuation">,</span> json<span class="token operator">=</span>formData<span class="token punctuation">)</span>
<span class="token keyword">print</span><span class="token punctuation">(</span>res<span class="token punctuation">)</span>
<span class="token keyword">print</span><span class="token punctuation">(</span>res<span class="token punctuation">.</span>text<span class="token punctuation">)</span>
</code></pre>
<h4>使用推理服务进行部署</h4>
<p>要在Kubernetes上部署模型，可以通过使用lightgbm和storageUri指定modelFormat来创建推理服务。<br> <code>旧框架</code></p>
<pre><code class="prism language-yaml"><span class="token key atrule">apiVersion</span><span class="token punctuation">:</span> <span class="token string">"serving.kserve.io/v1beta1"</span>
<span class="token key atrule">kind</span><span class="token punctuation">:</span> <span class="token string">"InferenceService"</span>
<span class="token key atrule">metadata</span><span class="token punctuation">:</span>
<span class="token key atrule">name</span><span class="token punctuation">:</span> <span class="token string">"lightgbm-iris"</span>
<span class="token key atrule">spec</span><span class="token punctuation">:</span>
<span class="token key atrule">predictor</span><span class="token punctuation">:</span>
<span class="token key atrule">lightgbm</span><span class="token punctuation">:</span>
<span class="token key atrule">storageUri</span><span class="token punctuation">:</span> <span class="token string">"gs://kfserving-examples/models/lightgbm/iris"</span>
</code></pre>
<p><code>新框架</code></p>
<pre><code class="prism language-yaml"><span class="token key atrule">apiVersion</span><span class="token punctuation">:</span> <span class="token string">"serving.kserve.io/v1beta1"</span>
<span class="token key atrule">kind</span><span class="token punctuation">:</span> <span class="token string">"InferenceService"</span>
<span class="token key atrule">metadata</span><span class="token punctuation">:</span>
<span class="token key atrule">name</span><span class="token punctuation">:</span> <span class="token string">"lightgbm-iris"</span>
<span class="token key atrule">spec</span><span class="token punctuation">:</span>
<span class="token key atrule">predictor</span><span class="token punctuation">:</span>
<span class="token key atrule">model</span><span class="token punctuation">:</span>
<span class="token key atrule">modelFormat</span><span class="token punctuation">:</span>
<span class="token key atrule">name</span><span class="token punctuation">:</span> lightgbm
<span class="token key atrule">storageUri</span><span class="token punctuation">:</span> <span class="token string">"gs://kfserving-examples/models/lightgbm/iris"</span>
</code></pre>
<p>应用上述yaml创建推理服务</p>
<pre><code>kubectl apply -f lightgbm.yaml
</code></pre>
<p><code>期望输出</code></p>
<pre><code>$ inferenceservice.serving.kserve.io/lightgbm-iris created
</code></pre>
<h4>测试已部署的模型</h4>
<p>要测试部署的模型，第一步是确定入口IP和端口，并设置INGRESS_HOST和INGRESS_PORT，然后运行以下curl命令将推理请求发送到推理服务。</p>
<pre><code class="prism language-powershell">MODEL_NAME=lightgbm-iris
INPUT_PATH=@<span class="token punctuation">.</span><span class="token operator">/</span>iris-input<span class="token punctuation">.</span>json
SERVICE_HOSTNAME=$<span class="token punctuation">(</span>kubectl get inferenceservice lightgbm-iris <span class="token operator">-</span>o jsonpath=<span class="token string">'{.status.url}'</span> <span class="token punctuation">|</span> cut <span class="token operator">-</span>d <span class="token string">"/"</span> <span class="token operator">-</span>f 3<span class="token punctuation">)</span>
curl <span class="token operator">-</span>v <span class="token operator">-</span>H <span class="token string">"Host: ${SERVICE_HOSTNAME}"</span> http:<span class="token operator">/</span><span class="token operator">/</span>$<span class="token punctuation">{</span>INGRESS_HOST<span class="token punctuation">}</span>:$<span class="token punctuation">{</span>INGRESS_PORT<span class="token punctuation">}</span><span class="token operator">/</span>v1/models/<span class="token variable">$MODEL_NAME</span>:predict <span class="token operator">-</span>d <span class="token variable">$INPUT_PATH</span>
</code></pre>
<p><code>期望输出</code></p>
<pre><code class="prism language-powershell"><span class="token operator">*</span> Trying 169<span class="token punctuation">.</span>63<span class="token punctuation">.</span>251<span class="token punctuation">.</span>68<span class="token punctuation">.</span><span class="token punctuation">.</span><span class="token punctuation">.</span>
<span class="token operator">*</span> TCP_NODELAY <span class="token function">set</span>
<span class="token operator">*</span> Connected to 169<span class="token punctuation">.</span>63<span class="token punctuation">.</span>251<span class="token punctuation">.</span>68 <span class="token punctuation">(</span>169<span class="token punctuation">.</span>63<span class="token punctuation">.</span>251<span class="token punctuation">.</span>68<span class="token punctuation">)</span> port 80 <span class="token punctuation">(</span><span class="token comment">#0)</span>
> POST <span class="token operator">/</span>models/lightgbm-iris:predict HTTP/1<span class="token punctuation">.</span>1
> Host: lightgbm-iris<span class="token punctuation">.</span>default<span class="token punctuation">.</span>svc<span class="token punctuation">.</span>cluster<span class="token punctuation">.</span>local
> User-Agent: curl/7<span class="token punctuation">.</span>60<span class="token punctuation">.</span>0
> Accept: <span class="token operator">*</span><span class="token operator">/</span><span class="token operator">*</span>
> Content-Length: 76
> Content-<span class="token function">Type</span>: application/x-www-form-urlencoded
>
<span class="token operator">*</span> upload completely sent off: 76 out of 76 bytes
< HTTP/1<span class="token punctuation">.</span>1 200 OK
< content-length: 27
< content-<span class="token function">type</span>: application/json<span class="token punctuation">;</span> charset=UTF-8
< date: Tue<span class="token punctuation">,</span> 21 May 2019 22:40:09 GMT
< server: istio-envoy
< x-envoy-upstream-service-time: 13032
<
<span class="token operator">*</span> Connection <span class="token comment">#0 to host 169.63.251.68 left intact</span>
<span class="token punctuation">{</span><span class="token string">"predictions"</span>: <span class="token punctuation">[</span><span class="token punctuation">[</span>0<span class="token punctuation">.</span>9<span class="token punctuation">,</span> 0<span class="token punctuation">.</span>05<span class="token punctuation">,</span> 0<span class="token punctuation">.</span>05<span class="token punctuation">]</span><span class="token punctuation">]</span><span class="token punctuation">}</span>
</code></pre>
<h3>使用V2协议部署模型</h3>
<h4>在本地测试模型</h4>
<p>一旦您获得了模型序列化的model.bst，我们就可以使用实现KServe V2推理协议的MLServer来启动本地服务器。有关MLServer的更多详细信息，请查看LightGBM示例文档。</p>
<p>要在本地运行MLServer，首先要在本地环境以及LightGBM运行时中安装mlserver包。</p>
<pre><code>pip install mlserver mlserver-lightgbm
</code></pre>
<p>下一步是提供模型设置，以便MLServer知道：</p>
<ul>
<li>为模型提供服务的推理运行时（即mlserver_lightgbm.LightGBMModel）</li>
<li>型号的名称和版本</li>
</ul>
<p>这些可以通过环境变量或创建本地model-settings.json文件来指定：</p>
<pre><code class="prism language-python"><span class="token punctuation">{</span>
<span class="token string">"name"</span><span class="token punctuation">:</span> <span class="token string">"lightgbm-iris"</span><span class="token punctuation">,</span>
<span class="token string">"version"</span><span class="token punctuation">:</span> <span class="token string">"v1.0.0"</span><span class="token punctuation">,</span>
<span class="token string">"implementation"</span><span class="token punctuation">:</span> <span class="token string">"mlserver_lightgbm.LightGBMModel"</span>
<span class="token punctuation">}</span>
</code></pre>
<p>有了本地安装的mlserver软件包和本地model-settings.json文件，您现在应该可以启动我们的服务器了：</p>
<pre><code>mlserver start .
</code></pre>
<h4>使用REST端点部署推理服务</h4>
<p>当您使用推理服务部署模型时，KServe会注入合理的默认值，这样它就可以开箱即用，而无需任何进一步的配置。但是，您仍然可以通过提供类似于本地文件的model-settings.json文件来覆盖这些默认值。您甚至可以提供一组model-settings.json文件来加载多个模型。</p>
<p>要使用V2推理协议部署LightGBM模型，需要将protocolVersion字段设置为V2。<br> <code>旧框架</code></p>
<pre><code class="prism language-yaml"><span class="token key atrule">apiVersion</span><span class="token punctuation">:</span> <span class="token string">"serving.kserve.io/v1beta1"</span>
<span class="token key atrule">kind</span><span class="token punctuation">:</span> <span class="token string">"InferenceService"</span>
<span class="token key atrule">metadata</span><span class="token punctuation">:</span>
<span class="token key atrule">name</span><span class="token punctuation">:</span> <span class="token string">"lightgbm-v2-iris"</span>
<span class="token key atrule">spec</span><span class="token punctuation">:</span>
<span class="token key atrule">predictor</span><span class="token punctuation">:</span>
<span class="token key atrule">lightgbm</span><span class="token punctuation">:</span>
<span class="token key atrule">protocolVersion</span><span class="token punctuation">:</span> v2
<span class="token key atrule">storageUri</span><span class="token punctuation">:</span> <span class="token string">"gs://kfserving-examples/models/lightgbm/v2/iris"</span>
</code></pre>
<p><code>新框架</code></p>
<pre><code class="prism language-yaml"><span class="token key atrule">apiVersion</span><span class="token punctuation">:</span> <span class="token string">"serving.kserve.io/v1beta1"</span>
<span class="token key atrule">kind</span><span class="token punctuation">:</span> <span class="token string">"InferenceService"</span>
<span class="token key atrule">metadata</span><span class="token punctuation">:</span>
<span class="token key atrule">name</span><span class="token punctuation">:</span> <span class="token string">"lightgbm-v2-iris"</span>
<span class="token key atrule">spec</span><span class="token punctuation">:</span>
<span class="token key atrule">predictor</span><span class="token punctuation">:</span>
<span class="token key atrule">model</span><span class="token punctuation">:</span>
<span class="token key atrule">modelFormat</span><span class="token punctuation">:</span>
<span class="token key atrule">name</span><span class="token punctuation">:</span> lightgbm
<span class="token key atrule">protocolVersion</span><span class="token punctuation">:</span> v2
<span class="token key atrule">storageUri</span><span class="token punctuation">:</span> <span class="token string">"gs://kfserving-examples/models/lightgbm/v2/iris"</span>
</code></pre>
<p>应用推理服务yaml来获取REST端点<br> <code>kubectl</code></p>
<pre><code>kubectl apply -f lightgbm-v2.yaml
</code></pre>
<p><code>期望输出</code></p>
<pre><code>$ inferenceservice.serving.kserve.io/lightgbm-v2-iris created
</code></pre>
<h4>用curl测试展开的模型</h4>
<p>现在，您可以通过发送示例请求来测试已部署的模型。</p>
<p>请注意，此请求需要遵循V2数据平面协议。您可以在下面看到一个负载示例：</p>
<pre><code class="prism language-python"><span class="token punctuation">{</span>
<span class="token string">"inputs"</span><span class="token punctuation">:</span> <span class="token punctuation">[</span>
<span class="token punctuation">{</span>
<span class="token string">"name"</span><span class="token punctuation">:</span> <span class="token string">"input-0"</span><span class="token punctuation">,</span>
<span class="token string">"shape"</span><span class="token punctuation">:</span> <span class="token punctuation">[</span><span class="token number">2</span><span class="token punctuation">,</span> <span class="token number">4</span><span class="token punctuation">]</span><span class="token punctuation">,</span>
<span class="token string">"datatype"</span><span class="token punctuation">:</span> <span class="token string">"FP32"</span><span class="token punctuation">,</span>
<span class="token string">"data"</span><span class="token punctuation">:</span> <span class="token punctuation">[</span>
<span class="token punctuation">[</span><span class="token number">6.8</span><span class="token punctuation">,</span> <span class="token number">2.8</span><span class="token punctuation">,</span> <span class="token number">4.8</span><span class="token punctuation">,</span> <span class="token number">1.4</span><span class="token punctuation">]</span><span class="token punctuation">,</span>
<span class="token punctuation">[</span><span class="token number">6.0</span><span class="token punctuation">,</span> <span class="token number">3.4</span><span class="token punctuation">,</span> <span class="token number">4.5</span><span class="token punctuation">,</span> <span class="token number">1.6</span><span class="token punctuation">]</span>
<span class="token punctuation">]</span>
<span class="token punctuation">}</span>
<span class="token punctuation">]</span>
<span class="token punctuation">}</span>
</code></pre>
<p>现在，假设您的入口可以在${INGRESS_HOST}: ${INGRESS_PORT}访问，或者您可以按照此说明查找入口IP和端口。</p>
<p>您可以使用curl将推理请求发送为：</p>
<pre><code class="prism language-powershell">SERVICE_HOSTNAME=$<span class="token punctuation">(</span>kubectl get inferenceservice lightgbm-v2-iris <span class="token operator">-</span>o jsonpath=<span class="token string">'{.status.url}'</span> <span class="token punctuation">|</span> cut <span class="token operator">-</span>d <span class="token string">"/"</span> <span class="token operator">-</span>f 3<span class="token punctuation">)</span>
curl <span class="token operator">-</span>v \
<span class="token operator">-</span>H <span class="token string">"Host: ${SERVICE_HOSTNAME}"</span> \
<span class="token operator">-</span>H <span class="token string">"Content-Type: application/json"</span> \
<span class="token operator">-</span>d @<span class="token punctuation">.</span><span class="token operator">/</span>iris-input-v2<span class="token punctuation">.</span>json \
http:<span class="token operator">/</span><span class="token operator">/</span>$<span class="token punctuation">{</span>INGRESS_HOST<span class="token punctuation">}</span>:$<span class="token punctuation">{</span>INGRESS_PORT<span class="token punctuation">}</span><span class="token operator">/</span>v2/models/lightgbm-v2-iris/infer
</code></pre>
<p><code>期望输出</code></p>
<pre><code class="prism language-python"><span class="token punctuation">{</span>
<span class="token string">"model_name"</span><span class="token punctuation">:</span><span class="token string">"lightgbm-v2-iris"</span><span class="token punctuation">,</span>
<span class="token string">"model_version"</span><span class="token punctuation">:</span>null<span class="token punctuation">,</span>
<span class="token string">"id"</span><span class="token punctuation">:</span><span class="token string">"96253e27-83cf-4262-b279-1bd4b18d7922"</span><span class="token punctuation">,</span>
<span class="token string">"parameters"</span><span class="token punctuation">:</span>null<span class="token punctuation">,</span>
<span class="token string">"outputs"</span><span class="token punctuation">:</span><span class="token punctuation">[</span>
<span class="token punctuation">{</span>
<span class="token string">"name"</span><span class="token punctuation">:</span><span class="token string">"predict"</span><span class="token punctuation">,</span>
<span class="token string">"shape"</span><span class="token punctuation">:</span><span class="token punctuation">[</span><span class="token number">2</span><span class="token punctuation">,</span><span class="token number">3</span><span class="token punctuation">]</span><span class="token punctuation">,</span>
<span class="token string">"datatype"</span><span class="token punctuation">:</span><span class="token string">"FP64"</span><span class="token punctuation">,</span>
<span class="token string">"parameters"</span><span class="token punctuation">:</span>null<span class="token punctuation">,</span>
<span class="token string">"data"</span><span class="token punctuation">:</span>
<span class="token punctuation">[</span><span class="token number">8.796664107010673e-06</span><span class="token punctuation">,</span><span class="token number">0.9992300031041593</span><span class="token punctuation">,</span><span class="token number">0.0007612002317336916</span><span class="token punctuation">,</span><span class="token number">4.974786820804187e-06</span><span class="token punctuation">,</span><span class="token number">0.9999919650711493</span><span class="token punctuation">,</span><span class="token number">3.0601420299625077e-06</span><span class="token punctuation">]</span>
<span class="token punctuation">}</span>
<span class="token punctuation">]</span>
<span class="token punctuation">}</span>
</code></pre>
<h4>使用gRPC端点创建推理服务</h4>
<p>创建推理服务yaml并公开gRPC端口，目前只允许一个端口公开HTTP或gRPC端口并且默认情况下公开HTTP端口。<br> <code>旧框架</code></p>
<pre><code class="prism language-yaml"><span class="token key atrule">apiVersion</span><span class="token punctuation">:</span> <span class="token string">"serving.kserve.io/v1beta1"</span>
<span class="token key atrule">kind</span><span class="token punctuation">:</span> <span class="token string">"InferenceService"</span>
<span class="token key atrule">metadata</span><span class="token punctuation">:</span>
<span class="token key atrule">name</span><span class="token punctuation">:</span> <span class="token string">"lightgbm-v2-iris"</span>
<span class="token key atrule">spec</span><span class="token punctuation">:</span>
<span class="token key atrule">predictor</span><span class="token punctuation">:</span>
<span class="token key atrule">lightgbm</span><span class="token punctuation">:</span>
<span class="token key atrule">protocolVersion</span><span class="token punctuation">:</span> v2
<span class="token key atrule">storageUri</span><span class="token punctuation">:</span> <span class="token string">"gs://kfserving-examples/models/lightgbm/v2/iris"</span>
<span class="token key atrule">ports</span><span class="token punctuation">:</span>
<span class="token punctuation">-</span> <span class="token key atrule">name</span><span class="token punctuation">:</span> h2c
<span class="token key atrule">protocol</span><span class="token punctuation">:</span> TCP
<span class="token key atrule">containerPort</span><span class="token punctuation">:</span> <span class="token number">9000</span>
</code></pre>
<p><code>新框架</code></p>
<pre><code class="prism language-yaml"><span class="token key atrule">apiVersion</span><span class="token punctuation">:</span> <span class="token string">"serving.kserve.io/v1beta1"</span>
<span class="token key atrule">kind</span><span class="token punctuation">:</span> <span class="token string">"InferenceService"</span>
<span class="token key atrule">metadata</span><span class="token punctuation">:</span>
<span class="token key atrule">name</span><span class="token punctuation">:</span> <span class="token string">"lightgbm-v2-iris"</span>
<span class="token key atrule">spec</span><span class="token punctuation">:</span>
<span class="token key atrule">predictor</span><span class="token punctuation">:</span>
<span class="token key atrule">model</span><span class="token punctuation">:</span>
<span class="token key atrule">modelFormat</span><span class="token punctuation">:</span>
<span class="token key atrule">name</span><span class="token punctuation">:</span> lightgbm
<span class="token key atrule">protocolVersion</span><span class="token punctuation">:</span> v2
<span class="token key atrule">storageUri</span><span class="token punctuation">:</span> <span class="token string">"gs://kfserving-examples/models/lightgbm/v2/iris"</span>
<span class="token key atrule">ports</span><span class="token punctuation">:</span>
<span class="token punctuation">-</span> <span class="token key atrule">name</span><span class="token punctuation">:</span> h2c
<span class="token key atrule">protocol</span><span class="token punctuation">:</span> TCP
<span class="token key atrule">containerPort</span><span class="token punctuation">:</span> <span class="token number">9000</span>
</code></pre>
<p>应用推理服务yaml获取gRPC端点<br> <code>kubectl</code></p>
<pre><code class="prism language-powershell">kubectl apply <span class="token operator">-</span>f lightgbm-v2-grpc<span class="token punctuation">.</span>yaml
</code></pre>
<h4>使用grpcurl测试已部署的模型</h4>
<p>gRPC推理服务就绪后，grpurl可用于向推理服务发送gRPC请求。</p>
<pre><code class="prism language-powershell"><span class="token comment"># download the proto file</span>
curl <span class="token operator">-</span>O https:<span class="token operator">/</span><span class="token operator">/</span>raw<span class="token punctuation">.</span>githubusercontent<span class="token punctuation">.</span>com/kserve/kserve/master/docs/predict-api/v2/grpc_predict_v2<span class="token punctuation">.</span>proto
INPUT_PATH=iris-input-v2-grpc<span class="token punctuation">.</span>json
PROTO_FILE=grpc_predict_v2<span class="token punctuation">.</span>proto
SERVICE_HOSTNAME=$<span class="token punctuation">(</span>kubectl get inferenceservice lightgbm-v2-iris <span class="token operator">-</span>o jsonpath=<span class="token string">'{.status.url}'</span> <span class="token punctuation">|</span> cut <span class="token operator">-</span>d <span class="token string">"/"</span> <span class="token operator">-</span>f 3<span class="token punctuation">)</span>
</code></pre>
<p>gRPC API遵循KServe预测V2协议。</p>
<p>例如，ServerReady API可用于检查服务器是否已就绪：</p>
<pre><code class="prism language-powershell">grpcurl \
<span class="token operator">-</span>plaintext \
<span class="token operator">-</span>proto $<span class="token punctuation">{</span>PROTO_FILE<span class="token punctuation">}</span> \
<span class="token operator">-</span>authority $<span class="token punctuation">{</span>SERVICE_HOSTNAME<span class="token punctuation">}</span>" \
$<span class="token punctuation">{</span>INGRESS_HOST<span class="token punctuation">}</span>:$<span class="token punctuation">{</span>INGRESS_PORT<span class="token punctuation">}</span> \
inference<span class="token punctuation">.</span>GRPCInferenceService<span class="token punctuation">.</span>ServerReady
</code></pre>
<p><code>期望输出</code></p>
<pre><code class="prism language-python"><span class="token punctuation">{</span>
<span class="token string">"ready"</span><span class="token punctuation">:</span> true
<span class="token punctuation">}</span>
</code></pre>
<p>ModelInfer API按照grpc_predict_v2.proto文件中定义的ModelInferRequest模式获取输入。请注意，输入文件与上一个curl示例中使用的文件不同。</p>
<pre><code class="prism language-powershell">grpcurl \
<span class="token operator">-</span>vv \
<span class="token operator">-</span>plaintext \
<span class="token operator">-</span>proto $<span class="token punctuation">{</span>PROTO_FILE<span class="token punctuation">}</span> \
<span class="token operator">-</span>authority $<span class="token punctuation">{</span>SERVICE_HOSTNAME<span class="token punctuation">}</span> \
<span class="token operator">-</span>d @ \
$<span class="token punctuation">{</span>INGRESS_HOST<span class="token punctuation">}</span>:$<span class="token punctuation">{</span>INGRESS_PORT<span class="token punctuation">}</span> \
inference<span class="token punctuation">.</span>GRPCInferenceService<span class="token punctuation">.</span>ModelInfer \
<<< $<span class="token punctuation">(</span><span class="token function">cat</span> <span class="token string">"<span class="token variable">$INPUT_PATH</span>"</span><span class="token punctuation">)</span>
</code></pre>
<p><code>期望输出</code></p>
<pre><code class="prism language-powershell">Resolved method descriptor:
<span class="token operator">/</span><span class="token operator">/</span> The ModelInfer API performs inference <span class="token keyword">using</span> the specified model<span class="token punctuation">.</span> Errors are
<span class="token operator">/</span><span class="token operator">/</span> indicated by the google<span class="token punctuation">.</span>rpc<span class="token punctuation">.</span>Status returned <span class="token keyword">for</span> the request<span class="token punctuation">.</span> The OK code
<span class="token operator">/</span><span class="token operator">/</span> indicates success and other codes indicate failure<span class="token punctuation">.</span>
rpc ModelInfer <span class="token punctuation">(</span> <span class="token punctuation">.</span>inference<span class="token punctuation">.</span>ModelInferRequest <span class="token punctuation">)</span> returns <span class="token punctuation">(</span> <span class="token punctuation">.</span>inference<span class="token punctuation">.</span>ModelInferResponse <span class="token punctuation">)</span><span class="token punctuation">;</span>
Request metadata to send:
<span class="token punctuation">(</span>empty<span class="token punctuation">)</span>
Response headers received:
accept-encoding: identity<span class="token punctuation">,</span>gzip
content-<span class="token function">type</span>: application/grpc
date: Sun<span class="token punctuation">,</span> 25 Sep 2022 10:25:05 GMT
grpc-accept-encoding: identity<span class="token punctuation">,</span>deflate<span class="token punctuation">,</span>gzip
server: istio-envoy
x-envoy-upstream-service-time: 99
Estimated response size: 91 bytes
Response contents:
<span class="token punctuation">{</span>
<span class="token string">"modelName"</span>: <span class="token string">"lightgbm-v2-iris"</span><span class="token punctuation">,</span>
<span class="token string">"outputs"</span>: <span class="token punctuation">[</span>
<span class="token punctuation">{</span>
<span class="token string">"name"</span>: <span class="token string">"predict"</span><span class="token punctuation">,</span>
<span class="token string">"datatype"</span>: <span class="token string">"FP64"</span><span class="token punctuation">,</span>
<span class="token string">"shape"</span>: <span class="token punctuation">[</span>
<span class="token string">"2"</span><span class="token punctuation">,</span>
<span class="token string">"3"</span>
<span class="token punctuation">]</span><span class="token punctuation">,</span>
<span class="token string">"contents"</span>: <span class="token punctuation">{</span>
<span class="token string">"fp64Contents"</span>: <span class="token punctuation">[</span>
8<span class="token punctuation">.</span>796664107010673e-06<span class="token punctuation">,</span>
0<span class="token punctuation">.</span>9992300031041593<span class="token punctuation">,</span>
0<span class="token punctuation">.</span>0007612002317336916<span class="token punctuation">,</span>
4<span class="token punctuation">.</span>974786820804187e-06<span class="token punctuation">,</span>
0<span class="token punctuation">.</span>9999919650711493<span class="token punctuation">,</span>
3<span class="token punctuation">.</span>0601420299625077e-06
<span class="token punctuation">]</span>
<span class="token punctuation">}</span>
<span class="token punctuation">}</span>
<span class="token punctuation">]</span>
<span class="token punctuation">}</span>
</code></pre>
<h2>使用推理服务部署Paddle模型</h2>
<p>在本例中，我们使用经过训练的paddle resnet50模型，通过运行带有paddle预测器的推理服务来对图像进行分类。</p>
<h3>创建推理服务</h3>
<p><code>旧框架</code></p>
<pre><code class="prism language-yaml"><span class="token key atrule">apiVersion</span><span class="token punctuation">:</span> <span class="token string">"serving.kserve.io/v1beta1"</span>
<span class="token key atrule">kind</span><span class="token punctuation">:</span> <span class="token string">"InferenceService"</span>
<span class="token key atrule">metadata</span><span class="token punctuation">:</span>
<span class="token key atrule">name</span><span class="token punctuation">:</span> <span class="token string">"paddle-resnet50"</span>
<span class="token key atrule">spec</span><span class="token punctuation">:</span>
<span class="token key atrule">predictor</span><span class="token punctuation">:</span>
<span class="token key atrule">paddle</span><span class="token punctuation">:</span>
<span class="token key atrule">storageUri</span><span class="token punctuation">:</span> <span class="token string">"https://zhouti-mcp-edge.cdn.bcebos.com/resnet50.tar.gz"</span>
</code></pre>
<p><code>新框架</code></p>
<pre><code class="prism language-yaml"><span class="token key atrule">apiVersion</span><span class="token punctuation">:</span> <span class="token string">"serving.kserve.io/v1beta1"</span>
<span class="token key atrule">kind</span><span class="token punctuation">:</span> <span class="token string">"InferenceService"</span>
<span class="token key atrule">metadata</span><span class="token punctuation">:</span>
<span class="token key atrule">name</span><span class="token punctuation">:</span> <span class="token string">"paddle-resnet50"</span>
<span class="token key atrule">spec</span><span class="token punctuation">:</span>
<span class="token key atrule">predictor</span><span class="token punctuation">:</span>
<span class="token key atrule">model</span><span class="token punctuation">:</span>
<span class="token key atrule">modelFormat</span><span class="token punctuation">:</span>
<span class="token key atrule">name</span><span class="token punctuation">:</span> paddle
<span class="token key atrule">storageUri</span><span class="token punctuation">:</span> <span class="token string">"https://zhouti-mcp-edge.cdn.bcebos.com/resnet50.tar.gz"</span>
</code></pre>
<p>应用上述yaml创建推理服务</p>
<pre><code class="prism language-powershell">kubectl apply <span class="token operator">-</span>f paddle<span class="token punctuation">.</span>yaml
</code></pre>
<p><code>期望输出</code></p>
<pre><code>$ inferenceservice.serving.kserve.io/paddle-resnet50 created
</code></pre>
<h3>运行预测</h3>
<p>第一步是确定入口IP和端口，并设置INGRESS_HOST和INGRESS_PORT</p>
<pre><code class="prism language-powershell">MODEL_NAME=paddle-resnet50
SERVICE_HOSTNAME=$<span class="token punctuation">(</span>kubectl get inferenceservice $<span class="token punctuation">{</span>MODEL_NAME<span class="token punctuation">}</span> <span class="token operator">-</span>o jsonpath=<span class="token string">'{.status.url}'</span> <span class="token punctuation">|</span> cut <span class="token operator">-</span>d <span class="token string">"/"</span> <span class="token operator">-</span>f 3<span class="token punctuation">)</span>
curl <span class="token operator">-</span>v <span class="token operator">-</span>H <span class="token string">"Host: ${SERVICE_HOSTNAME}"</span> http:<span class="token operator">/</span><span class="token operator">/</span>$<span class="token punctuation">{</span>INGRESS_HOST<span class="token punctuation">}</span>:$<span class="token punctuation">{</span>INGRESS_PORT<span class="token punctuation">}</span><span class="token operator">/</span>v1/models/$<span class="token punctuation">{</span>MODEL_NAME<span class="token punctuation">}</span>:predict <span class="token operator">-</span>d @<span class="token punctuation">.</span><span class="token operator">/</span>jay<span class="token punctuation">.</span>json
</code></pre>
<p><code>期望输出</code></p>
<pre><code class="prism language-python"><span class="token operator">*</span> Trying <span class="token number">127.0</span><span class="token number">.0</span><span class="token number">.1</span><span class="token punctuation">:</span><span class="token number">80.</span><span class="token punctuation">.</span><span class="token punctuation">.</span>
<span class="token operator">*</span> TCP_NODELAY <span class="token builtin">set</span>
<span class="token operator">*</span> Connected to localhost <span class="token punctuation">(</span><span class="token number">127.0</span><span class="token number">.0</span><span class="token number">.1</span><span class="token punctuation">)</span> port <span class="token number">80</span> <span class="token punctuation">(</span><span class="token comment">#0)</span>
<span class="token operator">></span> POST <span class="token operator">/</span>v1<span class="token operator">/</span>models<span class="token operator">/</span>paddle<span class="token operator">-</span>resnet50<span class="token punctuation">:</span>predict HTTP<span class="token operator">/</span><span class="token number">1.1</span>
<span class="token operator">></span> Host<span class="token punctuation">:</span> paddle<span class="token operator">-</span>resnet50<span class="token punctuation">.</span>default<span class="token punctuation">.</span>example<span class="token punctuation">.</span>com
<span class="token operator">></span> User<span class="token operator">-</span>Agent<span class="token punctuation">:</span> curl<span class="token operator">/</span><span class="token number">7.68</span><span class="token number">.0</span>
<span class="token operator">></span> Accept<span class="token punctuation">:</span> <span class="token operator">*</span><span class="token operator">/</span><span class="token operator">*</span>
<span class="token operator">></span> Content<span class="token operator">-</span>Length<span class="token punctuation">:</span> <span class="token number">3010209</span>
<span class="token operator">></span> Content<span class="token operator">-</span>Type<span class="token punctuation">:</span> application<span class="token operator">/</span>x<span class="token operator">-</span>www<span class="token operator">-</span>form<span class="token operator">-</span>urlencoded
<span class="token operator">></span> Expect<span class="token punctuation">:</span> <span class="token number">100</span><span class="token operator">-</span><span class="token keyword">continue</span>
<span class="token operator">></span>
<span class="token operator">*</span> Mark bundle <span class="token keyword">as</span> <span class="token keyword">not</span> supporting multiuse
<span class="token operator"><</span> HTTP<span class="token operator">/</span><span class="token number">1.1</span> <span class="token number">100</span> Continue
<span class="token operator">*</span> We are completely uploaded <span class="token keyword">and</span> fine
<span class="token operator">*</span> Mark bundle <span class="token keyword">as</span> <span class="token keyword">not</span> supporting multiuse
<span class="token operator"><</span> HTTP<span class="token operator">/</span><span class="token number">1.1</span> <span class="token number">200</span> OK
<span class="token operator"><</span> content<span class="token operator">-</span>length<span class="token punctuation">:</span> <span class="token number">23399</span>
<span class="token operator"><</span> content<span class="token operator">-</span><span class="token builtin">type</span><span class="token punctuation">:</span> application<span class="token operator">/</span>json<span class="token punctuation">;</span> charset<span class="token operator">=</span>UTF<span class="token operator">-</span><span class="token number">8</span>
<span class="token operator"><</span> date<span class="token punctuation">:</span> Mon<span class="token punctuation">,</span> <span class="token number">17</span> May <span class="token number">2021</span> <span class="token number">03</span><span class="token punctuation">:</span><span class="token number">34</span><span class="token punctuation">:</span><span class="token number">58</span> GMT
<span class="token operator"><</span> server<span class="token punctuation">:</span> istio<span class="token operator">-</span>envoy
<span class="token operator"><</span> x<span class="token operator">-</span>envoy<span class="token operator">-</span>upstream<span class="token operator">-</span>service<span class="token operator">-</span>time<span class="token punctuation">:</span> <span class="token number">511</span>
<span class="token operator"><</span>
<span class="token punctuation">{</span><span class="token string">"predictions"</span><span class="token punctuation">:</span> <span class="token punctuation">[</span><span class="token punctuation">[</span><span class="token number">6.736678770380422e-09</span><span class="token punctuation">,</span> <span class="token number">1.1535990829258935e-08</span><span class="token punctuation">,</span> <span class="token number">5.142250714129659e-08</span><span class="token punctuation">,</span> <span class="token number">6.647170636142619e-08</span><span class="token punctuation">,</span> <span class="token number">4.094492567219277e-08</span><span class="token punctuation">,</span> <span class="token number">1.3402451770616608e-07</span><span class="token punctuation">,</span> <span class="token number">9.355561303436843e-08</span><span class="token punctuation">,</span> <span class="token number">2.8935891904779965e-08</span><span class="token punctuation">,</span> <span class="token number">6.845367295227334e-08</span><span class="token punctuation">,</span> <span class="token number">7.680615965455218e-08</span><span class="token punctuation">,</span> <span class="token number">2.0334689452283783e-06</span><span class="token punctuation">,</span> <span class="token number">1.1085678579547675e-06</span><span class="token punctuation">,</span> <span class="token number">2.3477592492326949e-07</span><span class="token punctuation">,</span> <span class="token number">6.582037030966603e-07</span><span class="token punctuation">,</span> <span class="token number">0.00012373103527352214</span><span class="token punctuation">,</span> <span class="token number">4.2878804151769145e-07</span><span class="token punctuation">,</span> <span class="token number">6.419959845516132e-06</span><span class="token punctuation">,</span> <span class="token number">0.9993496537208557</span><span class="token punctuation">,</span> <span class="token number">7.372002437477931e-05</span><span class="token punctuation">,</span> <span class="token number">3.101135735050775e-05</span><span class="token punctuation">,</span> <span class="token number">5.6028093240456656e-06</span><span class="token punctuation">,</span> <span class="token number">2.1862508674530545e-06</span><span class="token punctuation">,</span> <span class="token number">1.9544044604913324e-08</span><span class="token punctuation">,</span> <span class="token number">3.728893887000595e-07</span><span class="token punctuation">,</span> <span class="token number">4.2903633357127546e-07</span><span class="token punctuation">,</span> <span class="token number">1.8251179767503345e-07</span><span class="token punctuation">,</span> <span class="token number">7.159925985433802e-08</span><span class="token punctuation">,</span> <span class="token number">9.231618136595898e-09</span><span class="token punctuation">,</span> <span class="token number">6.469241498052725e-07</span><span class="token punctuation">,</span> <span class="token number">7.031690341108288e-09</span><span class="token punctuation">,</span> <span class="token number">4.451231561120039e-08</span><span class="token punctuation">,</span> <span class="token number">1.2455971898361895e-07</span><span class="token punctuation">,</span> <span class="token number">9.44632745358831e-08</span><span class="token punctuation">,</span> <span class="token number">4.347704418705689e-08</span><span class="token punctuation">,</span> <span class="token number">4.658220120745682e-07</span><span class="token punctuation">,</span> <span class="token number">6.797721141538204e-08</span><span class="token punctuation">,</span> <span class="token number">2.1060276367279585e-07</span><span class="token punctuation">,</span> <span class="token number">2.2605123106700376e-08</span><span class="token punctuation">,</span> <span class="token number">1.4311490303953178e-07</span><span class="token punctuation">,</span> <span class="token number">7.951298641728499e-08</span><span class="token punctuation">,</span> <span class="token number">1.2341783417468832e-07</span><span class="token punctuation">,</span> <span class="token number">1.0921713737843675e-06</span><span class="token punctuation">,</span> <span class="token number">1.5243892448779661e-05</span><span class="token punctuation">,</span> <span class="token number">3.1173343018053856e-07</span><span class="token punctuation">,</span> <span class="token number">2.4152058131221565e-07</span><span class="token punctuation">,</span> <span class="token number">6.863762536113427e-08</span><span class="token punctuation">,</span> <span class="token number">8.467682022228473e-08</span><span class="token punctuation">,</span> <span class="token number">9.4246772164297e-08</span><span class="token punctuation">,</span> <span class="token number">1.0219210366813058e-08</span><span class="token punctuation">,</span> <span class="token number">3.3770753304906975e-08</span><span class="token punctuation">,</span> <span class="token number">3.6928835100979995e-08</span><span class="token punctuation">,</span> <span class="token number">1.3694031508748594e-07</span><span class="token punctuation">,</span> <span class="token number">1.0674284567357972e-07</span><span class="token punctuation">,</span> <span class="token number">2.599483650556067e-07</span><span class="token punctuation">,</span> <span class="token number">3.4866405940192635e-07</span><span class="token punctuation">,</span> <span class="token number">3.132053549848024e-08</span><span class="token punctuation">,</span> <span class="token number">3.574873232992104e-07</span><span class="token punctuation">,</span> <span class="token number">6.64843895492595e-08</span><span class="token punctuation">,</span> <span class="token number">3.1638955988455564e-07</span><span class="token punctuation">,</span> <span class="token number">1.2095878219042788e-06</span><span class="token punctuation">,</span> <span class="token number">8.66409024524728e-08</span><span class="token punctuation">,</span> <span class="token number">4.0144172430700564e-08</span><span class="token punctuation">,</span> <span class="token number">1.2544761318622477e-07</span><span class="token punctuation">,</span> <span class="token number">3.3201178695208e-08</span><span class="token punctuation">,</span> <span class="token number">1.9731444922399533e-07</span><span class="token punctuation">,</span> <span class="token number">3.806405572959193e-07</span><span class="token punctuation">,</span> <span class="token number">1.3827865075199952e-07</span><span class="token punctuation">,</span> <span class="token number">2.300225965257141e-08</span><span class="token punctuation">,</span> <span class="token number">7.14422512260171e-08</span><span class="token punctuation">,</span> <span class="token number">2.851114544455413e-08</span><span class="token punctuation">,</span> <span class="token number">2.982567437470607e-08</span><span class="token punctuation">,</span> <span class="token number">8.936032713791064e-08</span><span class="token punctuation">,</span> <span class="token number">6.22388370175031e-07</span><span class="token punctuation">,</span> <span class="token number">6.478838798784636e-08</span><span class="token punctuation">,</span> <span class="token number">1.3663023423760023e-07</span><span class="token punctuation">,</span> <span class="token number">9.973181391842445e-08</span><span class="token punctuation">,</span> <span class="token number">2.5761554667269593e-08</span><span class="token punctuation">,</span> <span class="token number">4.130220077058766e-08</span><span class="token punctuation">,</span> <span class="token number">3.9384463690339544e-08</span><span class="token punctuation">,</span> <span class="token number">1.2158079698565416e-07</span><span class="token punctuation">,</span> <span class="token number">4.302821707824478e-06</span><span class="token punctuation">,</span> <span class="token number">1.8179063090428826e-06</span><span class="token punctuation">,</span> <span class="token number">1.8520155435908237e-06</span><span class="token punctuation">,</span> <span class="token number">1.6246107179540559e-06</span><span class="token punctuation">,</span> <span class="token number">1.6448313544970006e-05</span><span class="token punctuation">,</span> <span class="token number">1.0544916221988387e-05</span><span class="token punctuation">,</span> <span class="token number">3.993061909568496e-06</span><span class="token punctuation">,</span> <span class="token number">2.646479799750523e-07</span><span class="token punctuation">,</span> <span class="token number">1.9193475964129902e-05</span><span class="token punctuation">,</span> <span class="token number">4.803242745765601e-07</span><span class="token punctuation">,</span> <span class="token number">1.696285067964709e-07</span><span class="token punctuation">,</span> <span class="token number">4.550505764200352e-06</span><span class="token punctuation">,</span> <span class="token number">4.235929372953251e-05</span><span class="token punctuation">,</span> <span class="token number">4.443338639248395e-06</span><span class="token punctuation">,</span> <span class="token number">5.104009687784128e-06</span><span class="token punctuation">,</span> <span class="token number">1.3506396498996764e-05</span><span class="token punctuation">,</span> <span class="token number">4.1758724478313525e-07</span><span class="token punctuation">,</span> <span class="token number">4.494491463447048e-07</span><span class="token punctuation">,</span> <span class="token number">3.156698369366495e-07</span><span class="token punctuation">,</span> <span class="token number">1.0557599807725637e-06</span><span class="token punctuation">,</span> <span class="token number">1.336463917311903e-08</span><span class="token punctuation">,</span> <span class="token number">1.3893659556174498e-08</span><span class="token punctuation">,</span> <span class="token number">6.770379457066156e-08</span><span class="token punctuation">,</span> <span class="token number">1.4129696523923485e-07</span><span class="token punctuation">,</span> <span class="token number">7.170518756538513e-08</span><span class="token punctuation">,</span> <span class="token number">7.934466594861078e-08</span><span class="token punctuation">,</span> <span class="token number">2.639154317307657e-08</span><span class="token punctuation">,</span> <span class="token number">2.6134321373660896e-08</span><span class="token punctuation">,</span> <span class="token number">7.196725881897237e-09</span><span class="token punctuation">,</span> <span class="token number">2.1752363466021052e-08</span><span class="token punctuation">,</span> <span class="token number">6.684639686227456e-08</span><span class="token punctuation">,</span> <span class="token number">3.417795824134373e-08</span><span class="token punctuation">,</span> <span class="token number">1.6228275967478112e-07</span><span class="token punctuation">,</span> <span class="token number">4.107114648377319e-07</span><span class="token punctuation">,</span> <span class="token number">6.472135396506928e-07</span><span class="token punctuation">,</span> <span class="token number">2.951379372007068e-07</span><span class="token punctuation">,</span> <span class="token number">5.653474133282543e-09</span><span class="token punctuation">,</span> <span class="token number">4.830144462175667e-08</span><span class="token punctuation">,</span> <span class="token number">8.887481861563629e-09</span><span class="token punctuation">,</span> <span class="token number">3.7306168820805397e-08</span><span class="token punctuation">,</span> <span class="token number">1.7784264727538357e-08</span><span class="token punctuation">,</span> <span class="token number">4.641905082536368e-09</span><span class="token punctuation">,</span> <span class="token number">3.413118676576232e-08</span><span class="token punctuation">,</span> <span class="token number">1.937393818707278e-07</span><span class="token punctuation">,</span> <span class="token number">1.2980176506971475e-06</span><span class="token punctuation">,</span> <span class="token number">3.5641004814124244e-08</span><span class="token punctuation">,</span> <span class="token number">2.149332445355867e-08</span><span class="token punctuation">,</span> <span class="token number">3.055293689158134e-07</span><span class="token punctuation">,</span> <span class="token number">1.5532516783878236e-07</span><span class="token punctuation">,</span> <span class="token number">1.4520978766086046e-06</span><span class="token punctuation">,</span> <span class="token number">3.488464628276233e-08</span><span class="token punctuation">,</span> <span class="token number">3.825438398052938e-05</span><span class="token punctuation">,</span> <span class="token number">4.5088432898410247e-07</span><span class="token punctuation">,</span> <span class="token number">4.1766969616219285e-07</span><span class="token punctuation">,</span> <span class="token number">6.770622462681786e-07</span><span class="token punctuation">,</span> <span class="token number">1.4142248971893423e-07</span><span class="token punctuation">,</span> <span class="token number">1.4235997696232516e-05</span><span class="token punctuation">,</span> <span class="token number">6.293820433711517e-07</span><span class="token punctuation">,</span> <span class="token number">4.762866865348769e-06</span><span class="token punctuation">,</span> <span class="token number">9.024900577969674e-07</span><span class="token punctuation">,</span> <span class="token number">9.058987870957935e-07</span><span class="token punctuation">,</span> <span class="token number">1.5713684433649178e-06</span><span class="token punctuation">,</span> <span class="token number">1.5720647184025438e-07</span><span class="token punctuation">,</span> <span class="token number">1.818536503606083e-07</span><span class="token punctuation">,</span> <span class="token number">7.193188622522939e-08</span><span class="token punctuation">,</span> <span class="token number">1.1952824934269302e-06</span><span class="token punctuation">,</span> <span class="token number">8.874837362782273e-07</span><span class="token punctuation">,</span> <span class="token number">2.0870831463071227e-07</span><span class="token punctuation">,</span> <span class="token number">9.906239029078279e-08</span><span class="token punctuation">,</span> <span class="token number">7.793621747964607e-09</span><span class="token punctuation">,</span> <span class="token number">1.0058498389753368e-07</span><span class="token punctuation">,</span> <span class="token number">4.2059440374941914e-07</span><span class="token punctuation">,</span> <span class="token number">1.843624630737395e-07</span><span class="token punctuation">,</span> <span class="token number">1.6437947181202617e-07</span><span class="token punctuation">,</span> <span class="token number">7.025352743994517e-08</span><span class="token punctuation">,</span> <span class="token number">2.570448600636155e-07</span><span class="token punctuation">,</span> <span class="token number">7.586877615040066e-08</span><span class="token punctuation">,</span> <span class="token number">7.841313731660193e-07</span><span class="token punctuation">,</span> <span class="token number">2.495309274763713e-07</span><span class="token punctuation">,</span> <span class="token number">5.157681925993529e-08</span><span class="token punctuation">,</span> <span class="token number">4.0674127177453556e-08</span><span class="token punctuation">,</span> <span class="token number">7.531796519799627e-09</span><span class="token punctuation">,</span> <span class="token number">4.797485431140558e-08</span><span class="token punctuation">,</span> <span class="token number">1.7419973019627832e-08</span><span class="token punctuation">,</span> <span class="token number">1.7958679165985814e-07</span><span class="token punctuation">,</span> <span class="token number">1.2566392371127222e-08</span><span class="token punctuation">,</span> <span class="token number">8.975440124459055e-08</span><span class="token punctuation">,</span> <span class="token number">3.26965476915575e-08</span><span class="token punctuation">,</span> <span class="token number">1.1208359751435637e-07</span><span class="token punctuation">,</span> <span class="token number">3.906746215420753e-08</span><span class="token punctuation">,</span> <span class="token number">4.6769045525252295e-08</span><span class="token punctuation">,</span> <span class="token number">1.8523553535487736e-07</span><span class="token punctuation">,</span> <span class="token number">1.4833052830454108e-07</span><span class="token punctuation">,</span> <span class="token number">1.2279349448363064e-07</span><span class="token punctuation">,</span> <span class="token number">1.0729105497375713e-06</span><span class="token punctuation">,</span> <span class="token number">3.6538490011395197e-09</span><span class="token punctuation">,</span> <span class="token number">1.6198403329781286e-07</span><span class="token punctuation">,</span> <span class="token number">1.6190719875908144e-08</span><span class="token punctuation">,</span> <span class="token number">1.2004933580556099e-07</span><span class="token punctuation">,</span> <span class="token number">1.4800277448046018e-08</span><span class="token punctuation">,</span> <span class="token number">4.02294837442696e-08</span><span class="token punctuation">,</span> <span class="token number">2.15060893538066e-07</span><span class="token punctuation">,</span> <span class="token number">1.1925696696835075e-07</span><span class="token punctuation">,</span> <span class="token number">4.8982514044837444e-08</span><span class="token punctuation">,</span> <span class="token number">7.608920071788816e-08</span><span class="token punctuation">,</span> <span class="token number">2.3137479487900237e-08</span><span class="token punctuation">,</span> <span class="token number">8.521050176568679e-08</span><span class="token punctuation">,</span> <span class="token number">9.586213423062873e-08</span><span class="token punctuation">,</span> <span class="token number">1.3351650807180704e-07</span><span class="token punctuation">,</span> <span class="token number">3.021699157557123e-08</span><span class="token punctuation">,</span> <span class="token number">4.423876376336011e-08</span><span class="token punctuation">,</span> <span class="token number">2.610667060309879e-08</span><span class="token punctuation">,</span> <span class="token number">2.3977091245797055e-07</span><span class="token punctuation">,</span> <span class="token number">1.3192564551900432e-07</span><span class="token punctuation">,</span> <span class="token number">1.6734931662654162e-08</span><span class="token punctuation">,</span> <span class="token number">1.588336999702733e-07</span><span class="token punctuation">,</span> <span class="token number">4.0643516285854275e-07</span><span class="token punctuation">,</span> <span class="token number">8.753454494581092e-08</span><span class="token punctuation">,</span> <span class="token number">8.366999395548191e-07</span><span class="token punctuation">,</span> <span class="token number">3.437598650180007e-08</span><span class="token punctuation">,</span> <span class="token number">7.847892646850596e-08</span><span class="token punctuation">,</span> <span class="token number">8.526394701391382e-09</span><span class="token punctuation">,</span> <span class="token number">9.601382799928615e-08</span><span class="token punctuation">,</span> <span class="token number">5.258924034023948e-07</span><span class="token punctuation">,</span> <span class="token number">1.3557448141909845e-07</span><span class="token punctuation">,</span> <span class="token number">1.0307226716577134e-07</span><span class="token punctuation">,</span> <span class="token number">1.0429813457335513e-08</span><span class="token punctuation">,</span> <span class="token number">5.187714435805901e-08</span><span class="token punctuation">,</span> <span class="token number">2.187001335585137e-08</span><span class="token punctuation">,</span> <span class="token number">1.1791439824548888e-08</span><span class="token punctuation">,</span> <span class="token number">2.98065643278278e-08</span><span class="token punctuation">,</span> <span class="token number">4.338393466696289e-08</span><span class="token punctuation">,</span> <span class="token number">2.9991046091026874e-08</span><span class="token punctuation">,</span> <span class="token number">2.8507610494443725e-08</span><span class="token punctuation">,</span> <span class="token number">3.058665143385042e-08</span><span class="token punctuation">,</span> <span class="token number">6.441099031917474e-08</span><span class="token punctuation">,</span> <span class="token number">1.5364101102477434e-08</span><span class="token punctuation">,</span> <span class="token number">1.5973883549236234e-08</span><span class="token punctuation">,</span> <span class="token number">2.5736850872704053e-08</span><span class="token punctuation">,</span> <span class="token number">1.0903765712555469e-07</span><span class="token punctuation">,</span> <span class="token number">3.2118737891551064e-08</span><span class="token punctuation">,</span> <span class="token number">6.819742992547617e-09</span><span class="token punctuation">,</span> <span class="token number">1.9251311300649832e-07</span><span class="token punctuation">,</span> <span class="token number">5.8258109447706374e-08</span><span class="token punctuation">,</span> <span class="token number">1.8765761922168167e-07</span><span class="token punctuation">,</span> <span class="token number">4.0070790419122204e-07</span><span class="token punctuation">,</span> <span class="token number">1.5791577823165426e-08</span><span class="token punctuation">,</span> <span class="token number">1.950158434738114e-07</span><span class="token punctuation">,</span> <span class="token number">1.0142063189277906e-08</span><span class="token punctuation">,</span> <span class="token number">2.744815041921811e-08</span><span class="token punctuation">,</span> <span class="token number">1.2843531571604672e-08</span><span class="token punctuation">,</span> <span class="token number">3.7297493094001766e-08</span><span class="token punctuation">,</span> <span class="token number">7.407496838141014e-08</span><span class="token punctuation">,</span> <span class="token number">4.20607833007125e-08</span><span class="token punctuation">,</span> <span class="token number">1.6924804668860816e-08</span><span class="token punctuation">,</span> <span class="token number">1.459203531339881e-07</span><span class="token punctuation">,</span> <span class="token number">4.344977000414474e-08</span><span class="token punctuation">,</span> <span class="token number">1.7191403856031684e-07</span><span class="token punctuation">,</span> <span class="token number">3.5817443233554513e-08</span><span class="token punctuation">,</span> <span class="token number">8.440249388286247e-09</span><span class="token punctuation">,</span> <span class="token number">4.194829728021432e-08</span><span class="token punctuation">,</span> <span class="token number">2.514032360068086e-08</span><span class="token punctuation">,</span> <span class="token number">2.8340199520471288e-08</span><span class="token punctuation">,</span> <span class="token number">8.747196034164517e-08</span><span class="token punctuation">,</span> <span class="token number">8.277125651545703e-09</span><span class="token punctuation">,</span> <span class="token number">1.1676293709683705e-08</span><span class="token punctuation">,</span> <span class="token number">1.4548514570833504e-07</span><span class="token punctuation">,</span> <span class="token number">7.200282148289716e-09</span><span class="token punctuation">,</span> <span class="token number">2.623600948936655e-06</span><span class="token punctuation">,</span> <span class="token number">5.675736929333652e-07</span><span class="token punctuation">,</span> <span class="token number">1.9483527466945816e-06</span><span class="token punctuation">,</span> <span class="token number">6.752595282932816e-08</span><span class="token punctuation">,</span> <span class="token number">8.168475318370838e-08</span><span class="token punctuation">,</span> <span class="token number">1.0933046468153407e-07</span><span class="token punctuation">,</span> <span class="token number">1.670913718498923e-07</span><span class="token punctuation">,</span> <span class="token number">3.1387276777650186e-08</span><span class="token punctuation">,</span> <span class="token number">2.973524537708272e-08</span><span class="token punctuation">,</span> <span class="token number">5.752163900751839e-08</span><span class="token punctuation">,</span> <span class="token number">5.850877471402782e-08</span><span class="token punctuation">,</span> <span class="token number">3.2544622285968217e-07</span><span class="token punctuation">,</span> <span class="token number">3.330221431951941e-08</span><span class="token punctuation">,</span> <span class="token number">4.186786668469722e-07</span><span class="token punctuation">,</span> <span class="token number">1.5085906568401697e-07</span><span class="token punctuation">,</span> <span class="token number">2.3346819943981245e-07</span><span class="token punctuation">,</span> <span class="token number">2.86402780602657e-07</span><span class="token punctuation">,</span> <span class="token number">2.2940319865938363e-07</span><span class="token punctuation">,</span> <span class="token number">1.8537603807544656e-07</span><span class="token punctuation">,</span> <span class="token number">3.151798182443599e-07</span><span class="token punctuation">,</span> <span class="token number">1.1075967449869495e-06</span><span class="token punctuation">,</span> <span class="token number">1.5369782602192572e-07</span><span class="token punctuation">,</span> <span class="token number">1.9237509718550427e-07</span><span class="token punctuation">,</span> <span class="token number">1.64044664074936e-07</span><span class="token punctuation">,</span> <span class="token number">2.900835340824415e-07</span><span class="token punctuation">,</span> <span class="token number">1.246654903752642e-07</span><span class="token punctuation">,</span> <span class="token number">5.802622027317739e-08</span><span class="token punctuation">,</span> <span class="token number">5.186220519703966e-08</span><span class="token punctuation">,</span> <span class="token number">6.0094205167615655e-09</span><span class="token punctuation">,</span> <span class="token number">1.2333241272699524e-07</span><span class="token punctuation">,</span> <span class="token number">1.3798474185477971e-07</span><span class="token punctuation">,</span> <span class="token number">1.7370231830682314e-07</span><span class="token punctuation">,</span> <span class="token number">5.617761189569137e-07</span><span class="token punctuation">,</span> <span class="token number">5.1604470030497396e-08</span><span class="token punctuation">,</span> <span class="token number">4.813277598714194e-08</span><span class="token punctuation">,</span> <span class="token number">8.032698417537176e-08</span><span class="token punctuation">,</span> <span class="token number">2.0645263703045202e-06</span><span class="token punctuation">,</span> <span class="token number">5.638597713186755e-07</span><span class="token punctuation">,</span> <span class="token number">8.794199857220519e-07</span><span class="token punctuation">,</span> <span class="token number">3.4785980460583232e-06</span><span class="token punctuation">,</span> <span class="token number">2.972389268052211e-07</span><span class="token punctuation">,</span> <span class="token number">3.3904532870110415e-07</span><span class="token punctuation">,</span> <span class="token number">9.469074058188198e-08</span><span class="token punctuation">,</span> <span class="token number">3.754845678827223e-08</span><span class="token punctuation">,</span> <span class="token number">1.5679037801419327e-07</span><span class="token punctuation">,</span> <span class="token number">8.203105039683578e-08</span><span class="token punctuation">,</span> <span class="token number">6.847962641387539e-09</span><span class="token punctuation">,</span> <span class="token number">1.8251624211984563e-08</span><span class="token punctuation">,</span> <span class="token number">6.050240841659615e-08</span><span class="token punctuation">,</span> <span class="token number">3.956342808919544e-08</span><span class="token punctuation">,</span> <span class="token number">1.0699947949888156e-07</span><span class="token punctuation">,</span> <span class="token number">3.2566634899922065e-07</span><span class="token punctuation">,</span> <span class="token number">3.5369430406717584e-07</span><span class="token punctuation">,</span> <span class="token number">7.326295303755614e-08</span><span class="token punctuation">,</span> <span class="token number">4.85765610847011e-07</span><span class="token punctuation">,</span> <span class="token number">7.717713401689252e-07</span><span class="token punctuation">,</span> <span class="token number">3.4567779749750116e-08</span><span class="token punctuation">,</span> <span class="token number">3.246204585138912e-07</span><span class="token punctuation">,</span> <span class="token number">3.1608601602783892e-06</span><span class="token punctuation">,</span> <span class="token number">5.33099466792919e-08</span><span class="token punctuation">,</span> <span class="token number">3.645687343123427e-07</span><span class="token punctuation">,</span> <span class="token number">5.48158936908294e-07</span><span class="token punctuation">,</span> <span class="token number">4.62306957160763e-08</span><span class="token punctuation">,</span> <span class="token number">1.3466177506415988e-07</span><span class="token punctuation">,</span> <span class="token number">4.3529482240955986e-08</span><span class="token punctuation">,</span> <span class="token number">1.6404105451783835e-07</span><span class="token punctuation">,</span> <span class="token number">2.463695381038633e-08</span><span class="token punctuation">,</span> <span class="token number">5.958712634424046e-08</span><span class="token punctuation">,</span> <span class="token number">9.493651020875404e-08</span><span class="token punctuation">,</span> <span class="token number">5.523462576206839e-08</span><span class="token punctuation">,</span> <span class="token number">5.7412357534758485e-08</span><span class="token punctuation">,</span> <span class="token number">1.1850350347231142e-05</span><span class="token punctuation">,</span> <span class="token number">5.8263944993086625e-06</span><span class="token punctuation">,</span> <span class="token number">7.4208674050169066e-06</span><span class="token punctuation">,</span> <span class="token number">9.127966222877149e-07</span><span class="token punctuation">,</span> <span class="token number">2.0019581370434025e-06</span><span class="token punctuation">,</span> <span class="token number">1.033498961078294e-06</span><span class="token punctuation">,</span> <span class="token number">3.5146850763112525e-08</span><span class="token punctuation">,</span> <span class="token number">2.058995278275688e-06</span><span class="token punctuation">,</span> <span class="token number">3.5655509122989315e-07</span><span class="token punctuation">,</span> <span class="token number">6.873234070781109e-08</span><span class="token punctuation">,</span> <span class="token number">2.1935298022413008e-09</span><span class="token punctuation">,</span> <span class="token number">5.560363547374436e-08</span><span class="token punctuation">,</span> <span class="token number">3.3266996979364194e-07</span><span class="token punctuation">,</span> <span class="token number">1.307369217329324e-07</span><span class="token punctuation">,</span> <span class="token number">2.718762992515167e-08</span><span class="token punctuation">,</span> <span class="token number">1.0462929189714032e-08</span><span class="token punctuation">,</span> <span class="token number">7.466680358447775e-07</span><span class="token punctuation">,</span> <span class="token number">6.923166040451179e-08</span><span class="token punctuation">,</span> <span class="token number">1.6145664361033596e-08</span><span class="token punctuation">,</span> <span class="token number">8.568521003837759e-09</span><span class="token punctuation">,</span> <span class="token number">4.76221018175238e-09</span><span class="token punctuation">,</span> <span class="token number">1.233977116044116e-07</span><span class="token punctuation">,</span> <span class="token number">8.340628632197422e-09</span><span class="token punctuation">,</span> <span class="token number">3.2649041248333788e-09</span><span class="token punctuation">,</span> <span class="token number">5.0632489312363305e-09</span><span class="token punctuation">,</span> <span class="token number">4.0704994930251814e-09</span><span class="token punctuation">,</span> <span class="token number">1.2043538610839732e-08</span><span class="token punctuation">,</span> <span class="token number">5.105608380517879e-09</span><span class="token punctuation">,</span> <span class="token number">7.267142887457112e-09</span><span class="token punctuation">,</span> <span class="token number">1.184516307262129e-07</span><span class="token punctuation">,</span> <span class="token number">7.53557927168913e-08</span><span class="token punctuation">,</span> <span class="token number">6.386964201965384e-08</span><span class="token punctuation">,</span> <span class="token number">1.6212936770898523e-08</span><span class="token punctuation">,</span> <span class="token number">2.610429419291904e-07</span><span class="token punctuation">,</span> <span class="token number">6.979425393183192e-07</span><span class="token punctuation">,</span> <span class="token number">6.647513117741255e-08</span><span class="token punctuation">,</span> <span class="token number">7.717492849224072e-07</span><span class="token punctuation">,</span> <span class="token number">6.651206945207377e-07</span><span class="token punctuation">,</span> <span class="token number">3.324495310152997e-07</span><span class="token punctuation">,</span> <span class="token number">3.707282019149716e-07</span><span class="token punctuation">,</span> <span class="token number">3.99564243025452e-07</span><span class="token punctuation">,</span> <span class="token number">6.411632114122767e-08</span><span class="token punctuation">,</span> <span class="token number">7.107352217872176e-08</span><span class="token punctuation">,</span> <span class="token number">1.6380016631956096e-07</span><span class="token punctuation">,</span> <span class="token number">6.876800995314625e-08</span><span class="token punctuation">,</span> <span class="token number">3.462474467141874e-07</span><span class="token punctuation">,</span> <span class="token number">2.0256503319160402e-07</span><span class="token punctuation">,</span> <span class="token number">6.19610148078209e-07</span><span class="token punctuation">,</span> <span class="token number">2.6841073363925716e-08</span><span class="token punctuation">,</span> <span class="token number">6.720335363752383e-07</span><span class="token punctuation">,</span> <span class="token number">1.1348340649419697e-06</span><span class="token punctuation">,</span> <span class="token number">1.8397931853542104e-06</span><span class="token punctuation">,</span> <span class="token number">6.397251581802266e-07</span><span class="token punctuation">,</span> <span class="token number">7.257533241045167e-08</span><span class="token punctuation">,</span> <span class="token number">4.2213909523525217e-07</span><span class="token punctuation">,</span> <span class="token number">3.9657925299252383e-07</span><span class="token punctuation">,</span> <span class="token number">1.4037439655112394e-07</span><span class="token punctuation">,</span> <span class="token number">3.249856774800719e-07</span><span class="token punctuation">,</span> <span class="token number">1.5857655455420172e-07</span><span class="token punctuation">,</span> <span class="token number">1.1122217102865761e-07</span><span class="token punctuation">,</span> <span class="token number">7.391420808744442e-08</span><span class="token punctuation">,</span> <span class="token number">3.42322238111592e-07</span><span class="token punctuation">,</span> <span class="token number">5.39796154441774e-08</span><span class="token punctuation">,</span> <span class="token number">8.517296379295658e-08</span><span class="token punctuation">,</span> <span class="token number">4.061009803990601e-06</span><span class="token punctuation">,</span> <span class="token number">1.4478755474556237e-05</span><span class="token punctuation">,</span> <span class="token number">7.317032757470088e-09</span><span class="token punctuation">,</span> <span class="token number">6.9484960008026064e-09</span><span class="token punctuation">,</span> <span class="token number">4.468917325084476e-08</span><span class="token punctuation">,</span> <span class="token number">9.23141172393116e-08</span><span class="token punctuation">,</span> <span class="token number">5.411982328951126e-08</span><span class="token punctuation">,</span> <span class="token number">2.2242811326123046e-07</span><span class="token punctuation">,</span> <span class="token number">1.7609554703312824e-08</span><span class="token punctuation">,</span> <span class="token number">2.0906279374344194e-08</span><span class="token punctuation">,</span> <span class="token number">3.6797682678724186e-09</span><span class="token punctuation">,</span> <span class="token number">6.177919686933819e-08</span><span class="token punctuation">,</span> <span class="token number">1.7920288541972695e-07</span><span class="token punctuation">,</span> <span class="token number">2.6279179721200308e-08</span><span class="token punctuation">,</span> <span class="token number">2.6988200119149042e-08</span><span class="token punctuation">,</span> <span class="token number">1.6432807115052128e-07</span><span class="token punctuation">,</span> <span class="token number">1.2827612749788386e-07</span><span class="token punctuation">,</span> <span class="token number">4.468908798571647e-08</span><span class="token punctuation">,</span> <span class="token number">6.316552969565237e-08</span><span class="token punctuation">,</span> <span class="token number">1.9461760203398626e-08</span><span class="token punctuation">,</span> <span class="token number">2.087125849925542e-08</span><span class="token punctuation">,</span> <span class="token number">2.2414580413965268e-08</span><span class="token punctuation">,</span> <span class="token number">2.4765244077684656e-08</span><span class="token punctuation">,</span> <span class="token number">6.785398465325443e-09</span><span class="token punctuation">,</span> <span class="token number">2.4248794971981624e-08</span><span class="token punctuation">,</span> <span class="token number">4.554979504689527e-09</span><span class="token punctuation">,</span> <span class="token number">2.8977037658250993e-08</span><span class="token punctuation">,</span> <span class="token number">2.0402325162649504e-08</span><span class="token punctuation">,</span> <span class="token number">1.600950270130852e-07</span><span class="token punctuation">,</span> <span class="token number">2.0199709638291097e-07</span><span class="token punctuation">,</span> <span class="token number">1.611188515937556e-08</span><span class="token punctuation">,</span> <span class="token number">5.964113825029926e-08</span><span class="token punctuation">,</span> <span class="token number">4.098318573397819e-09</span><span class="token punctuation">,</span> <span class="token number">3.9080127578472457e-08</span><span class="token punctuation">,</span> <span class="token number">7.511338218080255e-09</span><span class="token punctuation">,</span> <span class="token number">5.965624154669058e-07</span><span class="token punctuation">,</span> <span class="token number">1.6478223585636442e-07</span><span class="token punctuation">,</span> <span class="token number">1.4106989354445432e-08</span><span class="token punctuation">,</span> <span class="token number">3.2855584919389e-08</span><span class="token punctuation">,</span> <span class="token number">3.3387166364917675e-09</span><span class="token punctuation">,</span> <span class="token number">1.220043444050134e-08</span><span class="token punctuation">,</span> <span class="token number">4.624639160510924e-08</span><span class="token punctuation">,</span> <span class="token number">6.842309385746148e-09</span><span class="token punctuation">,</span> <span class="token number">1.74262879681919e-08</span><span class="token punctuation">,</span> <span class="token number">4.6611329906909305e-08</span><span class="token punctuation">,</span> <span class="token number">9.331947836699328e-08</span><span class="token punctuation">,</span> <span class="token number">1.2306078644996887e-07</span><span class="token punctuation">,</span> <span class="token number">1.2359445022980253e-08</span><span class="token punctuation">,</span> <span class="token number">1.1173199254699284e-08</span><span class="token punctuation">,</span> <span class="token number">2.7724862405875683e-08</span><span class="token punctuation">,</span> <span class="token number">2.419210147763806e-07</span><span class="token punctuation">,</span> <span class="token number">3.451186785241589e-07</span><span class="token punctuation">,</span> <span class="token number">2.593766978975509e-08</span><span class="token punctuation">,</span> <span class="token number">9.964568192799561e-08</span><span class="token punctuation">,</span> <span class="token number">9.797809674694236e-09</span><span class="token punctuation">,</span> <span class="token number">1.9085564417764544e-07</span><span class="token punctuation">,</span> <span class="token number">3.972706252852731e-08</span><span class="token punctuation">,</span> <span class="token number">2.6639204619982593e-08</span><span class="token punctuation">,</span> <span class="token number">6.874148805735558e-09</span><span class="token punctuation">,</span> <span class="token number">3.146993776681484e-08</span><span class="token punctuation">,</span> <span class="token number">2.4086594407890516e-07</span><span class="token punctuation">,</span> <span class="token number">1.3126927456141857e-07</span><span class="token punctuation">,</span> <span class="token number">2.1254339799270383e-07</span><span class="token punctuation">,</span> <span class="token number">2.050203384840188e-08</span><span class="token punctuation">,</span> <span class="token number">3.694976058454813e-08</span><span class="token punctuation">,</span> <span class="token number">6.563175816154398e-07</span><span class="token punctuation">,</span> <span class="token number">2.560050127442537e-08</span><span class="token punctuation">,</span> <span class="token number">2.6882981174480847e-08</span><span class="token punctuation">,</span> <span class="token number">6.880636078676616e-07</span><span class="token punctuation">,</span> <span class="token number">2.0092733166166e-07</span><span class="token punctuation">,</span> <span class="token number">2.788039665801989e-08</span><span class="token punctuation">,</span> <span class="token number">2.628409134786125e-08</span><span class="token punctuation">,</span> <span class="token number">5.1678345158734373e-08</span><span class="token punctuation">,</span> <span class="token number">1.8935413947929192e-07</span><span class="token punctuation">,</span> <span class="token number">4.61852835087484e-07</span><span class="token punctuation">,</span> <span class="token number">1.1086777718105623e-08</span><span class="token punctuation">,</span> <span class="token number">1.4542604276357451e-07</span><span class="token punctuation">,</span> <span class="token number">2.8737009216683873e-08</span><span class="token punctuation">,</span> <span class="token number">6.105167926762078e-07</span><span class="token punctuation">,</span> <span class="token number">1.2016463379893594e-08</span><span class="token punctuation">,</span> <span class="token number">1.3944705301582871e-07</span><span class="token punctuation">,</span> <span class="token number">2.093712758721722e-08</span><span class="token punctuation">,</span> <span class="token number">4.3801410498645055e-08</span><span class="token punctuation">,</span> <span class="token number">1.966320795077081e-08</span><span class="token punctuation">,</span> <span class="token number">6.654448991838535e-09</span><span class="token punctuation">,</span> <span class="token number">1.1149590584125235e-08</span><span class="token punctuation">,</span> <span class="token number">6.424939158478082e-08</span><span class="token punctuation">,</span> <span class="token number">6.971554888934861e-09</span><span class="token punctuation">,</span> <span class="token number">3.260019587614238e-09</span><span class="token punctuation">,</span> <span class="token number">1.4260189473702667e-08</span><span class="token punctuation">,</span> <span class="token number">2.7895078247297533e-08</span><span class="token punctuation">,</span> <span class="token number">8.11578289017234e-08</span><span class="token punctuation">,</span> <span class="token number">2.5995715802196173e-08</span><span class="token punctuation">,</span> <span class="token number">2.2855578762914774e-08</span><span class="token punctuation">,</span> <span class="token number">1.055962854934478e-07</span><span class="token punctuation">,</span> <span class="token number">8.145542551574181e-08</span><span class="token punctuation">,</span> <span class="token number">3.7793686402665116e-08</span><span class="token punctuation">,</span> <span class="token number">4.881891513264236e-08</span><span class="token punctuation">,</span> <span class="token number">2.342062366267328e-08</span><span class="token punctuation">,</span> <span class="token number">1.059935517133681e-08</span><span class="token punctuation">,</span> <span class="token number">3.604105103249822e-08</span><span class="token punctuation">,</span> <span class="token number">5.062430830093945e-08</span><span class="token punctuation">,</span> <span class="token number">3.6804440384230475e-08</span><span class="token punctuation">,</span> <span class="token number">1.501580193519203e-09</span><span class="token punctuation">,</span> <span class="token number">1.4475033367489232e-06</span><span class="token punctuation">,</span> <span class="token number">1.076210423889279e-06</span><span class="token punctuation">,</span> <span class="token number">1.304991315009829e-07</span><span class="token punctuation">,</span> <span class="token number">3.073601462233455e-08</span><span class="token punctuation">,</span> <span class="token number">1.7184021317007137e-08</span><span class="token punctuation">,</span> <span class="token number">2.0421090596300928e-08</span><span class="token punctuation">,</span> <span class="token number">7.904992216367646e-09</span><span class="token punctuation">,</span> <span class="token number">1.6902052379919041e-07</span><span class="token punctuation">,</span> <span class="token number">1.2416506933732308e-08</span><span class="token punctuation">,</span> <span class="token number">5.4758292122869534e-08</span><span class="token punctuation">,</span> <span class="token number">2.6250422280327257e-08</span><span class="token punctuation">,</span> <span class="token number">1.3261367115546818e-08</span><span class="token punctuation">,</span> <span class="token number">6.29807459517906e-08</span><span class="token punctuation">,</span> <span class="token number">1.270998595259698e-08</span><span class="token punctuation">,</span> <span class="token number">2.0171681569536304e-07</span><span class="token punctuation">,</span> <span class="token number">4.386637186826192e-08</span><span class="token punctuation">,</span> <span class="token number">6.962349630157405e-08</span><span class="token punctuation">,</span> <span class="token number">2.9565120485131047e-07</span><span class="token punctuation">,</span> <span class="token number">7.925131626507209e-07</span><span class="token punctuation">,</span> <span class="token number">2.0868920103112032e-07</span><span class="token punctuation">,</span> <span class="token number">1.7341794489311724e-07</span><span class="token punctuation">,</span> <span class="token number">4.2942417621816276e-08</span><span class="token punctuation">,</span> <span class="token number">4.213406956665722e-09</span><span class="token punctuation">,</span> <span class="token number">8.824785169281313e-08</span><span class="token punctuation">,</span> <span class="token number">1.7341569957807224e-08</span><span class="token punctuation">,</span> <span class="token number">7.321587247588468e-08</span><span class="token punctuation">,</span> <span class="token number">1.7941774288487977e-08</span><span class="token punctuation">,</span> <span class="token number">1.1245148101579616e-07</span><span class="token punctuation">,</span> <span class="token number">4.242405395871174e-07</span><span class="token punctuation">,</span> <span class="token number">8.259573469615589e-09</span><span class="token punctuation">,</span> <span class="token number">1.1336403105133286e-07</span><span class="token punctuation">,</span> <span class="token number">8.268798978861014e-08</span><span class="token punctuation">,</span> <span class="token number">2.2186977588489754e-08</span><span class="token punctuation">,</span> <span class="token number">1.9539720952366224e-08</span><span class="token punctuation">,</span> <span class="token number">1.0675703876472653e-08</span><span class="token punctuation">,</span> <span class="token number">3.288517547161973e-08</span><span class="token punctuation">,</span> <span class="token number">2.4340963022950746e-08</span><span class="token punctuation">,</span> <span class="token number">6.639137239972115e-08</span><span class="token punctuation">,</span> <span class="token number">5.604687380866835e-09</span><span class="token punctuation">,</span> <span class="token number">1.386604697728444e-08</span><span class="token punctuation">,</span> <span class="token number">6.675873720496384e-08</span><span class="token punctuation">,</span> <span class="token number">1.1355886009312144e-08</span><span class="token punctuation">,</span> <span class="token number">3.132159633878473e-07</span><span class="token punctuation">,</span> <span class="token number">3.12451788886392e-08</span><span class="token punctuation">,</span> <span class="token number">1.502181845580708e-07</span><span class="token punctuation">,</span> <span class="token number">1.3461754377885882e-08</span><span class="token punctuation">,</span> <span class="token number">1.8882955998833495e-07</span><span class="token punctuation">,</span> <span class="token number">4.645742279762999e-08</span><span class="token punctuation">,</span> <span class="token number">4.6453880742092224e-08</span><span class="token punctuation">,</span> <span class="token number">7.714453964524637e-09</span><span class="token punctuation">,</span> <span class="token number">3.5857155467056145e-08</span><span class="token punctuation">,</span> <span class="token number">7.60832108426257e-09</span><span class="token punctuation">,</span> <span class="token number">4.221501370693659e-08</span><span class="token punctuation">,</span> <span class="token number">4.3407251126836854e-09</span><span class="token punctuation">,</span> <span class="token number">1.340157496088068e-08</span><span class="token punctuation">,</span> <span class="token number">8.565600495558101e-08</span><span class="token punctuation">,</span> <span class="token number">1.7045413969185574e-08</span><span class="token punctuation">,</span> <span class="token number">5.4221903411644234e-08</span><span class="token punctuation">,</span> <span class="token number">3.021912675649219e-08</span><span class="token punctuation">,</span> <span class="token number">6.153376119755194e-08</span><span class="token punctuation">,</span> <span class="token number">3.938857240370908e-09</span><span class="token punctuation">,</span> <span class="token number">4.135628017820636e-08</span><span class="token punctuation">,</span> <span class="token number">1.781920389021252e-08</span><span class="token punctuation">,</span> <span class="token number">4.3105885083605244e-08</span><span class="token punctuation">,</span> <span class="token number">3.903354972578654e-09</span><span class="token punctuation">,</span> <span class="token number">7.663085455078544e-08</span><span class="token punctuation">,</span> <span class="token number">1.1890405993142394e-08</span><span class="token punctuation">,</span> <span class="token number">9.304217840622186e-09</span><span class="token punctuation">,</span> <span class="token number">1.0968062014171664e-09</span><span class="token punctuation">,</span> <span class="token number">1.0536767902635802e-08</span><span class="token punctuation">,</span> <span class="token number">1.1516804221400889e-07</span><span class="token punctuation">,</span> <span class="token number">8.134522886393825e-07</span><span class="token punctuation">,</span> <span class="token number">5.952623993721318e-08</span><span class="token punctuation">,</span> <span class="token number">2.806350174466843e-08</span><span class="token punctuation">,</span> <span class="token number">1.2833099027886874e-08</span><span class="token punctuation">,</span> <span class="token number">1.0605690192733164e-07</span><span class="token punctuation">,</span> <span class="token number">7.872949936427176e-07</span><span class="token punctuation">,</span> <span class="token number">2.7501393162765453e-08</span><span class="token punctuation">,</span> <span class="token number">3.936289072470345e-09</span><span class="token punctuation">,</span> <span class="token number">2.0519442145428002e-08</span><span class="token punctuation">,</span> <span class="token number">7.394815870753746e-09</span><span class="token punctuation">,</span> <span class="token number">3.598397313453461e-08</span><span class="token punctuation">,</span> <span class="token number">2.5378517065632877e-08</span><span class="token punctuation">,</span> <span class="token number">4.698972233541099e-08</span><span class="token punctuation">,</span> <span class="token number">7.54952989012736e-09</span><span class="token punctuation">,</span> <span class="token number">6.322805461422831e-07</span><span class="token punctuation">,</span> <span class="token number">5.582006412652163e-09</span><span class="token punctuation">,</span> <span class="token number">1.29640980617296e-07</span><span class="token punctuation">,</span> <span class="token number">1.5874988434916304e-08</span><span class="token punctuation">,</span> <span class="token number">3.3837810775594335e-08</span><span class="token punctuation">,</span> <span class="token number">6.474512037613067e-09</span><span class="token punctuation">,</span> <span class="token number">9.121148281110436e-08</span><span class="token punctuation">,</span> <span class="token number">1.3918511676536127e-08</span><span class="token punctuation">,</span> <span class="token number">8.230025549949005e-09</span><span class="token punctuation">,</span> <span class="token number">2.7061290097663004e-08</span><span class="token punctuation">,</span> <span class="token number">2.6095918315149902e-08</span><span class="token punctuation">,</span> <span class="token number">5.722363471960534e-09</span><span class="token punctuation">,</span> <span class="token number">6.963475698285038e-07</span><span class="token punctuation">,</span> <span class="token number">4.685091781198025e-08</span><span class="token punctuation">,</span> <span class="token number">9.590579885809802e-09</span><span class="token punctuation">,</span> <span class="token number">2.099205858030473e-07</span><span class="token punctuation">,</span> <span class="token number">3.082160660028421e-08</span><span class="token punctuation">,</span> <span class="token number">3.563162565001221e-08</span><span class="token punctuation">,</span> <span class="token number">7.326312925215461e-07</span><span class="token punctuation">,</span> <span class="token number">2.1759731225756695e-06</span><span class="token punctuation">,</span> <span class="token number">2.407518309155421e-07</span><span class="token punctuation">,</span> <span class="token number">2.974515780351794e-07</span><span class="token punctuation">,</span> <span class="token number">2.529018416908002e-08</span><span class="token punctuation">,</span> <span class="token number">7.667950718825978e-09</span><span class="token punctuation">,</span> <span class="token number">2.663289251358947e-07</span><span class="token punctuation">,</span> <span class="token number">3.4358880185436647e-08</span><span class="token punctuation">,</span> <span class="token number">2.3130198201215535e-08</span><span class="token punctuation">,</span> <span class="token number">3.1239693498719134e-08</span><span class="token punctuation">,</span> <span class="token number">2.8691621878351725e-07</span><span class="token punctuation">,</span> <span class="token number">3.895845068768722e-08</span><span class="token punctuation">,</span> <span class="token number">2.4184130253956937e-08</span><span class="token punctuation">,</span> <span class="token number">1.1582445225144511e-08</span><span class="token punctuation">,</span> <span class="token number">5.1545349322168477e-08</span><span class="token punctuation">,</span> <span class="token number">2.034345492063494e-08</span><span class="token punctuation">,</span> <span class="token number">8.201963197507212e-08</span><span class="token punctuation">,</span> <span class="token number">1.164153573540716e-08</span><span class="token punctuation">,</span> <span class="token number">5.496356720868789e-07</span><span class="token punctuation">,</span> <span class="token number">1.1682151246361627e-08</span><span class="token punctuation">,</span> <span class="token number">4.7576914852243135e-08</span><span class="token punctuation">,</span> <span class="token number">1.6349824605299546e-08</span><span class="token punctuation">,</span> <span class="token number">4.090862759653646e-08</span><span class="token punctuation">,</span> <span class="token number">2.1271189609706198e-07</span><span class="token punctuation">,</span> <span class="token number">1.6697286753242224e-07</span><span class="token punctuation">,</span> <span class="token number">3.989708119433999e-08</span><span class="token punctuation">,</span> <span class="token number">2.852450279533514e-06</span><span class="token punctuation">,</span> <span class="token number">1.2500372292834072e-07</span><span class="token punctuation">,</span> <span class="token number">2.4846613655427063e-07</span><span class="token punctuation">,</span> <span class="token number">1.245429093188477e-08</span><span class="token punctuation">,</span> <span class="token number">2.9700272463628608e-08</span><span class="token punctuation">,</span> <span class="token number">4.250991558762962e-09</span><span class="token punctuation">,</span> <span class="token number">1.61443480806156e-07</span><span class="token punctuation">,</span> <span class="token number">2.6386018703306036e-07</span><span class="token punctuation">,</span> <span class="token number">7.638056409575711e-09</span><span class="token punctuation">,</span> <span class="token number">3.4455793773702226e-09</span><span class="token punctuation">,</span> <span class="token number">7.273289526210647e-08</span><span class="token punctuation">,</span> <span class="token number">1.7631434090503717e-08</span><span class="token punctuation">,</span> <span class="token number">7.58661311550668e-09</span><span class="token punctuation">,</span> <span class="token number">2.1547013062672704e-08</span><span class="token punctuation">,</span> <span class="token number">1.2675349125856883e-07</span><span class="token punctuation">,</span> <span class="token number">2.5637149292379036e-08</span><span class="token punctuation">,</span> <span class="token number">3.500976220038865e-08</span><span class="token punctuation">,</span> <span class="token number">6.472243541111311e-08</span><span class="token punctuation">,</span> <span class="token number">8.387915251262257e-09</span><span class="token punctuation">,</span> <span class="token number">3.069512288789156e-08</span><span class="token punctuation">,</span> <span class="token number">7.520387867998579e-08</span><span class="token punctuation">,</span> <span class="token number">1.5724964441687916e-07</span><span class="token punctuation">,</span> <span class="token number">1.9634005354873807e-07</span><span class="token punctuation">,</span> <span class="token number">1.2290831818972947e-07</span><span class="token punctuation">,</span> <span class="token number">1.112118730439704e-09</span><span class="token punctuation">,</span> <span class="token number">1.546895944670723e-08</span><span class="token punctuation">,</span> <span class="token number">9.91701032404535e-09</span><span class="token punctuation">,</span> <span class="token number">6.882473257974198e-07</span><span class="token punctuation">,</span> <span class="token number">8.267616635748709e-08</span><span class="token punctuation">,</span> <span class="token number">4.469531234008173e-08</span><span class="token punctuation">,</span> <span class="token number">2.075201344098332e-08</span><span class="token punctuation">,</span> <span class="token number">8.649378457903367e-08</span><span class="token punctuation">,</span> <span class="token number">5.202766573120243e-08</span><span class="token punctuation">,</span> <span class="token number">4.5564942041664835e-08</span><span class="token punctuation">,</span> <span class="token number">2.0319955496006514e-08</span><span class="token punctuation">,</span> <span class="token number">8.705182352741758e-09</span><span class="token punctuation">,</span> <span class="token number">6.452066969586667e-08</span><span class="token punctuation">,</span> <span class="token number">2.1777438519166026e-08</span><span class="token punctuation">,</span> <span class="token number">1.030954166481024e-08</span><span class="token punctuation">,</span> <span class="token number">3.211904342492744e-08</span><span class="token punctuation">,</span> <span class="token number">2.3336936294526822e-07</span><span class="token punctuation">,</span> <span class="token number">8.054096056753224e-09</span><span class="token punctuation">,</span> <span class="token number">1.9623354319264763e-07</span><span class="token punctuation">,</span> <span class="token number">1.2888089884199871e-07</span><span class="token punctuation">,</span> <span class="token number">1.5392496166555247e-08</span><span class="token punctuation">,</span> <span class="token number">1.401903038100727e-09</span><span class="token punctuation">,</span> <span class="token number">5.696818305978013e-08</span><span class="token punctuation">,</span> <span class="token number">6.080025372057207e-09</span><span class="token punctuation">,</span> <span class="token number">1.0782793324892737e-08</span><span class="token punctuation">,</span> <span class="token number">2.4260730313585555e-08</span><span class="token punctuation">,</span> <span class="token number">1.9388659566743627e-08</span><span class="token punctuation">,</span> <span class="token number">2.2970310453729326e-07</span><span class="token punctuation">,</span> <span class="token number">1.9971754028347277e-08</span><span class="token punctuation">,</span> <span class="token number">2.8477993296860404e-08</span><span class="token punctuation">,</span> <span class="token number">5.2273552597625894e-08</span><span class="token punctuation">,</span> <span class="token number">2.7392806600801123e-07</span><span class="token punctuation">,</span> <span class="token number">9.857291161097237e-08</span><span class="token punctuation">,</span> <span class="token number">3.12910977129377e-08</span><span class="token punctuation">,</span> <span class="token number">4.151442212219081e-08</span><span class="token punctuation">,</span> <span class="token number">5.251196366629074e-09</span><span class="token punctuation">,</span> <span class="token number">1.580681100676884e-06</span><span class="token punctuation">,</span> <span class="token number">8.547603442821128e-07</span><span class="token punctuation">,</span> <span class="token number">1.068913135782168e-08</span><span class="token punctuation">,</span> <span class="token number">1.0621830597301596e-06</span><span class="token punctuation">,</span> <span class="token number">7.737313012512459e-08</span><span class="token punctuation">,</span> <span class="token number">6.394216711669287e-08</span><span class="token punctuation">,</span> <span class="token number">1.1698345758759388e-07</span><span class="token punctuation">,</span> <span class="token number">1.0486609625104393e-07</span><span class="token punctuation">,</span> <span class="token number">2.1161000063329993e-07</span><span class="token punctuation">,</span> <span class="token number">1.53396815250062e-08</span><span class="token punctuation">,</span> <span class="token number">5.094453570109181e-08</span><span class="token punctuation">,</span> <span class="token number">1.4005379966874898e-08</span><span class="token punctuation">,</span> <span class="token number">2.6282036102998063e-08</span><span class="token punctuation">,</span> <span class="token number">8.778433624456738e-08</span><span class="token punctuation">,</span> <span class="token number">7.772066545896905e-09</span><span class="token punctuation">,</span> <span class="token number">4.228875383205377e-08</span><span class="token punctuation">,</span> <span class="token number">3.3243779284930497e-07</span><span class="token punctuation">,</span> <span class="token number">7.729244799747903e-08</span><span class="token punctuation">,</span> <span class="token number">7.636901111496286e-10</span><span class="token punctuation">,</span> <span class="token number">5.989500806435899e-08</span><span class="token punctuation">,</span> <span class="token number">1.326090597331131e-07</span><span class="token punctuation">,</span> <span class="token number">1.2853634245857393e-07</span><span class="token punctuation">,</span> <span class="token number">8.844242671557367e-09</span><span class="token punctuation">,</span> <span class="token number">1.0194374766570036e-07</span><span class="token punctuation">,</span> <span class="token number">2.493779334145074e-07</span><span class="token punctuation">,</span> <span class="token number">1.6547971881664125e-07</span><span class="token punctuation">,</span> <span class="token number">1.1762754326127833e-08</span><span class="token punctuation">,</span> <span class="token number">1.1496195639892903e-07</span><span class="token punctuation">,</span> <span class="token number">2.9342709240154363e-07</span><span class="token punctuation">,</span> <span class="token number">1.326124099421122e-08</span><span class="token punctuation">,</span> <span class="token number">8.630262726683213e-08</span><span class="token punctuation">,</span> <span class="token number">5.7394842656322e-08</span><span class="token punctuation">,</span> <span class="token number">1.1094081031615133e-07</span><span class="token punctuation">,</span> <span class="token number">2.2933713239581266e-07</span><span class="token punctuation">,</span> <span class="token number">3.4706170026765903e-07</span><span class="token punctuation">,</span> <span class="token number">1.4751107357824367e-07</span><span class="token punctuation">,</span> <span class="token number">1.502495017291494e-08</span><span class="token punctuation">,</span> <span class="token number">6.454319390059027e-08</span><span class="token punctuation">,</span> <span class="token number">5.164533689594464e-08</span><span class="token punctuation">,</span> <span class="token number">6.23741556182722e-08</span><span class="token punctuation">,</span> <span class="token number">1.293601457064142e-07</span><span class="token punctuation">,</span> <span class="token number">1.4052071506398534e-08</span><span class="token punctuation">,</span> <span class="token number">5.386946000385251e-08</span><span class="token punctuation">,</span> <span class="token number">2.0827554791935654e-08</span><span class="token punctuation">,</span> <span class="token number">1.3040637902861363e-08</span><span class="token punctuation">,</span> <span class="token number">1.0578981601838677e-07</span><span class="token punctuation">,</span> <span class="token number">1.5079727688771527e-08</span><span class="token punctuation">,</span> <span class="token number">8.92632726845477e-07</span><span class="token punctuation">,</span> <span class="token number">4.6374381668101705e-08</span><span class="token punctuation">,</span> <span class="token number">7.481006036869076e-07</span><span class="token punctuation">,</span> <span class="token number">5.883147302654379e-09</span><span class="token punctuation">,</span> <span class="token number">2.8707685117979054e-09</span><span class="token punctuation">,</span> <span class="token number">8.381598490814213e-07</span><span class="token punctuation">,</span> <span class="token number">7.341958596640552e-09</span><span class="token punctuation">,</span> <span class="token number">1.4245998158912698e-08</span><span class="token punctuation">,</span> <span class="token number">1.0926417104428765e-07</span><span class="token punctuation">,</span> <span class="token number">1.1308178216040687e-07</span><span class="token punctuation">,</span> <span class="token number">2.52339901862797e-07</span><span class="token punctuation">,</span> <span class="token number">1.1782835684925885e-07</span><span class="token punctuation">,</span> <span class="token number">4.6678056975224536e-08</span><span class="token punctuation">,</span> <span class="token number">2.7959197179683315e-09</span><span class="token punctuation">,</span> <span class="token number">3.4363861090014325e-08</span><span class="token punctuation">,</span> <span class="token number">1.4674496640054713e-07</span><span class="token punctuation">,</span> <span class="token number">3.5396915620822256e-08</span><span class="token punctuation">,</span> <span class="token number">2.0581127557761647e-07</span><span class="token punctuation">,</span> <span class="token number">7.18387909159901e-08</span><span class="token punctuation">,</span> <span class="token number">2.7693943138729082e-08</span><span class="token punctuation">,</span> <span class="token number">4.5493386835460115e-08</span><span class="token punctuation">,</span> <span class="token number">1.9559182717898693e-08</span><span class="token punctuation">,</span> <span class="token number">1.5359708172013598e-08</span><span class="token punctuation">,</span> <span class="token number">1.2336623278486059e-08</span><span class="token punctuation">,</span> <span class="token number">2.9570605519779747e-08</span><span class="token punctuation">,</span> <span class="token number">2.877552560676122e-07</span><span class="token punctuation">,</span> <span class="token number">9.051845495378075e-07</span><span class="token punctuation">,</span> <span class="token number">2.3732602016934834e-07</span><span class="token punctuation">,</span> <span class="token number">1.6521676471370483e-08</span><span class="token punctuation">,</span> <span class="token number">1.5478875070584763e-08</span><span class="token punctuation">,</span> <span class="token number">3.526786329643983e-08</span><span class="token punctuation">,</span> <span class="token number">3.616410637619083e-08</span><span class="token punctuation">,</span> <span class="token number">1.61590953950963e-08</span><span class="token punctuation">,</span> <span class="token number">7.65007328595857e-08</span><span class="token punctuation">,</span> <span class="token number">1.9661483108279754e-08</span><span class="token punctuation">,</span> <span class="token number">4.917534823789538e-08</span><span class="token punctuation">,</span> <span class="token number">1.1712612746350715e-07</span><span class="token punctuation">,</span> <span class="token number">1.0889253054813253e-08</span><span class="token punctuation">,</span> <span class="token number">1.494120169809321e-06</span><span class="token punctuation">,</span> <span class="token number">1.018585660261806e-08</span><span class="token punctuation">,</span> <span class="token number">3.7575969003000864e-08</span><span class="token punctuation">,</span> <span class="token number">2.097097784314883e-08</span><span class="token punctuation">,</span> <span class="token number">3.368558054717141e-08</span><span class="token punctuation">,</span> <span class="token number">4.845588819080149e-09</span><span class="token punctuation">,</span> <span class="token number">6.039624622644624e-07</span><span class="token punctuation">,</span> <span class="token number">1.037331109898787e-08</span><span class="token punctuation">,</span> <span class="token number">2.841650257323636e-07</span><span class="token punctuation">,</span> <span class="token number">4.4990630954089283e-07</span><span class="token punctuation">,</span> <span class="token number">3.463186004637464e-08</span><span class="token punctuation">,</span> <span class="token number">7.720684180867465e-08</span><span class="token punctuation">,</span> <span class="token number">1.471122175189521e-07</span><span class="token punctuation">,</span> <span class="token number">1.1601575522490748e-07</span><span class="token punctuation">,</span> <span class="token number">4.007488030310924e-07</span><span class="token punctuation">,</span> <span class="token number">3.025649775167949e-08</span><span class="token punctuation">,</span> <span class="token number">6.706784461130155e-08</span><span class="token punctuation">,</span> <span class="token number">2.0128741340386114e-08</span><span class="token punctuation">,</span> <span class="token number">1.5987744461654074e-09</span><span class="token punctuation">,</span> <span class="token number">4.1919822280078733e-08</span><span class="token punctuation">,</span> <span class="token number">1.3167154477855547e-08</span><span class="token punctuation">,</span> <span class="token number">3.231814815762846e-08</span><span class="token punctuation">,</span> <span class="token number">9.247659704669786e-08</span><span class="token punctuation">,</span> <span class="token number">1.3075300842047e-07</span><span class="token punctuation">,</span> <span class="token number">1.0574301256838226e-07</span><span class="token punctuation">,</span> <span class="token number">3.762165334819656e-08</span><span class="token punctuation">,</span> <span class="token number">1.0942246575496029e-07</span><span class="token punctuation">,</span> <span class="token number">7.001474955359299e-08</span><span class="token punctuation">,</span> <span class="token number">2.742706151082075e-08</span><span class="token punctuation">,</span> <span class="token number">2.0766625752344225e-08</span><span class="token punctuation">,</span> <span class="token number">4.5403403703403455e-08</span><span class="token punctuation">,</span> <span class="token number">3.39040298058535e-08</span><span class="token punctuation">,</span> <span class="token number">1.0469661759771043e-07</span><span class="token punctuation">,</span> <span class="token number">2.8271578855765256e-08</span><span class="token punctuation">,</span> <span class="token number">3.406226767310727e-07</span><span class="token punctuation">,</span> <span class="token number">5.146206945028098e-07</span><span class="token punctuation">,</span> <span class="token number">6.740708613506285e-07</span><span class="token punctuation">,</span> <span class="token number">6.382248063374618e-09</span><span class="token punctuation">,</span> <span class="token number">3.63878704945364e-08</span><span class="token punctuation">,</span> <span class="token number">3.626059807970705e-08</span><span class="token punctuation">,</span> <span class="token number">1.6065602892467723e-07</span><span class="token punctuation">,</span> <span class="token number">3.639055989879125e-07</span><span class="token punctuation">,</span> <span class="token number">6.232691696084203e-09</span><span class="token punctuation">,</span> <span class="token number">4.805490050330263e-08</span><span class="token punctuation">,</span> <span class="token number">3.372633727849461e-08</span><span class="token punctuation">,</span> <span class="token number">6.328880317596486e-07</span><span class="token punctuation">,</span> <span class="token number">6.480631498106959e-08</span><span class="token punctuation">,</span> <span class="token number">2.1165197949812864e-07</span><span class="token punctuation">,</span> <span class="token number">8.38779143919055e-08</span><span class="token punctuation">,</span> <span class="token number">1.7589144363228115e-08</span><span class="token punctuation">,</span> <span class="token number">2.729027670511641e-09</span><span class="token punctuation">,</span> <span class="token number">2.144795097080987e-08</span><span class="token punctuation">,</span> <span class="token number">7.861271456022223e-08</span><span class="token punctuation">,</span> <span class="token number">2.0118186228046397e-08</span><span class="token punctuation">,</span> <span class="token number">2.8407685093156942e-08</span><span class="token punctuation">,</span> <span class="token number">2.4922530883486615e-07</span><span class="token punctuation">,</span> <span class="token number">2.0156670998972004e-08</span><span class="token punctuation">,</span> <span class="token number">2.6551649767725394e-08</span><span class="token punctuation">,</span> <span class="token number">2.7848242822869906e-08</span><span class="token punctuation">,</span> <span class="token number">6.907123761834555e-09</span><span class="token punctuation">,</span> <span class="token number">1.880543720744754e-08</span><span class="token punctuation">,</span> <span class="token number">1.3006903998302732e-08</span><span class="token punctuation">,</span> <span class="token number">3.685918272822164e-07</span><span class="token punctuation">,</span> <span class="token number">3.967941211158177e-07</span><span class="token punctuation">,</span> <span class="token number">2.7592133022835696e-08</span><span class="token punctuation">,</span> <span class="token number">2.5228947819755376e-08</span><span class="token punctuation">,</span> <span class="token number">1.547002881352455e-07</span><span class="token punctuation">,</span> <span class="token number">3.689306637966183e-08</span><span class="token punctuation">,</span> <span class="token number">1.440177199718562e-09</span><span class="token punctuation">,</span> <span class="token number">2.1504929392790473e-08</span><span class="token punctuation">,</span> <span class="token number">5.068111263994979e-08</span><span class="token punctuation">,</span> <span class="token number">5.081711407228795e-08</span><span class="token punctuation">,</span> <span class="token number">1.171875219085905e-08</span><span class="token punctuation">,</span> <span class="token number">5.409278358570191e-08</span><span class="token punctuation">,</span> <span class="token number">7.138276600926474e-07</span><span class="token punctuation">,</span> <span class="token number">2.5237213208129106e-07</span><span class="token punctuation">,</span> <span class="token number">7.072044638789521e-08</span><span class="token punctuation">,</span> <span class="token number">7.199763984999663e-08</span><span class="token punctuation">,</span> <span class="token number">1.2525473103153217e-08</span><span class="token punctuation">,</span> <span class="token number">3.4803417747752974e-07</span><span class="token punctuation">,</span> <span class="token number">1.9591827538079087e-07</span><span class="token punctuation">,</span> <span class="token number">1.2404700555634918e-07</span><span class="token punctuation">,</span> <span class="token number">1.234617457157583e-07</span><span class="token punctuation">,</span> <span class="token number">1.9201337408958352e-08</span><span class="token punctuation">,</span> <span class="token number">1.9895249181445251e-07</span><span class="token punctuation">,</span> <span class="token number">3.7876677794201896e-08</span><span class="token punctuation">,</span> <span class="token number">1.0629785052174157e-08</span><span class="token punctuation">,</span> <span class="token number">1.2437127772102485e-08</span><span class="token punctuation">,</span> <span class="token number">2.1861892207653e-07</span><span class="token punctuation">,</span> <span class="token number">2.6181456291851646e-07</span><span class="token punctuation">,</span> <span class="token number">1.112900775979142e-07</span><span class="token punctuation">,</span> <span class="token number">1.0776630432474121e-07</span><span class="token punctuation">,</span> <span class="token number">6.380325157095967e-09</span><span class="token punctuation">,</span> <span class="token number">3.895085143312826e-09</span><span class="token punctuation">,</span> <span class="token number">1.5762756788717525e-07</span><span class="token punctuation">,</span> <span class="token number">2.909027019271093e-09</span><span class="token punctuation">,</span> <span class="token number">1.0381050685737137e-08</span><span class="token punctuation">,</span> <span class="token number">2.8135211493918177e-08</span><span class="token punctuation">,</span> <span class="token number">1.0778002490496874e-08</span><span class="token punctuation">,</span> <span class="token number">1.3605974125141529e-08</span><span class="token punctuation">,</span> <span class="token number">2.9236465692861202e-08</span><span class="token punctuation">,</span> <span class="token number">1.9189795352758665e-07</span><span class="token punctuation">,</span> <span class="token number">2.199506354827463e-07</span><span class="token punctuation">,</span> <span class="token number">1.326399790002597e-08</span><span class="token punctuation">,</span> <span class="token number">4.9004846403022384e-08</span><span class="token punctuation">,</span> <span class="token number">2.980837132682268e-09</span><span class="token punctuation">,</span> <span class="token number">8.926045680368588e-09</span><span class="token punctuation">,</span> <span class="token number">1.0996975774446582e-08</span><span class="token punctuation">,</span> <span class="token number">7.71560149104289e-09</span><span class="token punctuation">,</span> <span class="token number">7.454491246505768e-09</span><span class="token punctuation">,</span> <span class="token number">5.086162246925596e-08</span><span class="token punctuation">,</span> <span class="token number">1.5129764108223753e-07</span><span class="token punctuation">,</span> <span class="token number">1.1960075596562092e-08</span><span class="token punctuation">,</span> <span class="token number">1.1323334270230134e-08</span><span class="token punctuation">,</span> <span class="token number">9.391332156383214e-09</span><span class="token punctuation">,</span> <span class="token number">9.585701832293125e-08</span><span class="token punctuation">,</span> <span class="token number">1.905532798218701e-08</span><span class="token punctuation">,</span> <span class="token number">1.8105303922766325e-08</span><span class="token punctuation">,</span> <span class="token number">6.179227796110354e-08</span><span class="token punctuation">,</span> <span class="token number">6.389401363549041e-08</span><span class="token punctuation">,</span> <span class="token number">1.1853179771037503e-08</span><span class="token punctuation">,</span> <span class="token number">9.37277544466042e-09</span><span class="token punctuation">,</span> <span class="token number">1.2332148457971925e-07</span><span class="token punctuation">,</span> <span class="token number">1.6522022860954166e-08</span><span class="token punctuation">,</span> <span class="token number">1.246116454467483e-07</span><span class="token punctuation">,</span> <span class="token number">4.196171854431441e-09</span><span class="token punctuation">,</span> <span class="token number">3.996593278543514e-08</span><span class="token punctuation">,</span> <span class="token number">1.2554556505506298e-08</span><span class="token punctuation">,</span> <span class="token number">1.4302138140465104e-08</span><span class="token punctuation">,</span> <span class="token number">6.631793780798034e-09</span><span class="token punctuation">,</span> <span class="token number">5.964224669696705e-09</span><span class="token punctuation">,</span> <span class="token number">5.556936244488497e-09</span><span class="token punctuation">,</span> <span class="token number">1.4192455921602232e-07</span><span class="token punctuation">,</span> <span class="token number">1.7613080771639034e-08</span><span class="token punctuation">,</span> <span class="token number">3.380189639301534e-07</span><span class="token punctuation">,</span> <span class="token number">7.85651934620546e-08</span><span class="token punctuation">,</span> <span class="token number">2.966783085867064e-08</span><span class="token punctuation">,</span> <span class="token number">2.8992105853831163e-06</span><span class="token punctuation">,</span> <span class="token number">1.3787366697215475e-06</span><span class="token punctuation">,</span> <span class="token number">5.313622430946907e-09</span><span class="token punctuation">,</span> <span class="token number">2.512852859126724e-08</span><span class="token punctuation">,</span> <span class="token number">8.406627216572815e-08</span><span class="token punctuation">,</span> <span class="token number">4.492839167369311e-08</span><span class="token punctuation">,</span> <span class="token number">5.408793057881667e-08</span><span class="token punctuation">,</span> <span class="token number">2.4239175999696272e-08</span><span class="token punctuation">,</span> <span class="token number">4.016805235096399e-07</span><span class="token punctuation">,</span> <span class="token number">4.1083545454512205e-08</span><span class="token punctuation">,</span> <span class="token number">5.4153481698904216e-08</span><span class="token punctuation">,</span> <span class="token number">8.640767212853007e-09</span><span class="token punctuation">,</span> <span class="token number">5.773256717134245e-08</span><span class="token punctuation">,</span> <span class="token number">2.6443152023603034e-07</span><span class="token punctuation">,</span> <span class="token number">8.953217047746875e-07</span><span class="token punctuation">,</span> <span class="token number">2.7994001783326894e-08</span><span class="token punctuation">,</span> <span class="token number">5.889480014786841e-09</span><span class="token punctuation">,</span> <span class="token number">4.1788819515886644e-08</span><span class="token punctuation">,</span> <span class="token number">2.8880645430717777e-08</span><span class="token punctuation">,</span> <span class="token number">2.135752907861388e-08</span><span class="token punctuation">,</span> <span class="token number">2.3024175277441827e-07</span><span class="token punctuation">,</span> <span class="token number">8.786625471657317e-08</span><span class="token punctuation">,</span> <span class="token number">2.0697297209437693e-09</span><span class="token punctuation">,</span> <span class="token number">2.236410523437371e-08</span><span class="token punctuation">,</span> <span class="token number">3.203276310870251e-09</span><span class="token punctuation">,</span> <span class="token number">1.176874686592555e-08</span><span class="token punctuation">,</span> <span class="token number">6.963571053120177e-08</span><span class="token punctuation">,</span> <span class="token number">2.271932153519174e-08</span><span class="token punctuation">,</span> <span class="token number">7.360382525689602e-09</span><span class="token punctuation">,</span> <span class="token number">6.922528772435044e-09</span><span class="token punctuation">,</span> <span class="token number">3.213871480056696e-08</span><span class="token punctuation">,</span> <span class="token number">1.370577820125618e-07</span><span class="token punctuation">,</span> <span class="token number">1.9815049157045905e-08</span><span class="token punctuation">,</span> <span class="token number">1.0578956377571558e-08</span><span class="token punctuation">,</span> <span class="token number">2.7049420481262132e-08</span><span class="token punctuation">,</span> <span class="token number">2.9755937713815683e-09</span><span class="token punctuation">,</span> <span class="token number">2.1773699288019088e-08</span><span class="token punctuation">,</span> <span class="token number">1.09755387001087e-08</span><span class="token punctuation">,</span> <span class="token number">1.991872444762066e-08</span><span class="token punctuation">,</span> <span class="token number">2.3882098076910552e-08</span><span class="token punctuation">,</span> <span class="token number">2.1357365653784655e-08</span><span class="token punctuation">,</span> <span class="token number">6.109098560358461e-09</span><span class="token punctuation">,</span> <span class="token number">1.1890497475519624e-08</span><span class="token punctuation">,</span> <span class="token number">1.1459891702259029e-08</span><span class="token punctuation">,</span> <span class="token number">3.73173456580389e-08</span><span class="token punctuation">,</span> <span class="token number">1.572620256240498e-08</span><span class="token punctuation">,</span> <span class="token number">3.404023374287135e-08</span><span class="token punctuation">,</span> <span class="token number">3.6921580459647885e-08</span><span class="token punctuation">,</span> <span class="token number">9.281765045443535e-08</span><span class="token punctuation">,</span> <span class="token number">1.2323201303843234e-07</span><span class="token punctuation">,</span> <span class="token number">4.2347593876002065e-08</span><span class="token punctuation">,</span> <span class="token number">1.7423728237986325e-08</span><span class="token punctuation">,</span> <span class="token number">5.8113389656000436e-08</span><span class="token punctuation">,</span> <span class="token number">3.931436154402945e-08</span><span class="token punctuation">,</span> <span class="token number">2.3690461148362374e-08</span><span class="token punctuation">,</span> <span class="token number">1.792850135018398e-08</span><span class="token punctuation">,</span> <span class="token number">1.440664210150544e-08</span><span class="token punctuation">,</span> <span class="token number">7.019830494670032e-09</span><span class="token punctuation">,</span> <span class="token number">6.041522482291839e-08</span><span class="token punctuation">,</span> <span class="token number">4.867479930226182e-08</span><span class="token punctuation">,</span> <span class="token number">1.0685319296044327e-08</span><span class="token punctuation">,</span> <span class="token number">1.0051243393149889e-08</span><span class="token punctuation">,</span> <span class="token number">4.2426261614991745e-08</span><span class="token punctuation">,</span> <span class="token number">2.607815297039906e-08</span><span class="token punctuation">,</span> <span class="token number">5.136670200300</span><span class="token operator">*</span> Connection <span class="token comment">#0 to host localhost left intact</span>
<span class="token number">841e-09</span><span class="token punctuation">,</span> <span class="token number">1.69729952315123e-09</span><span class="token punctuation">,</span> <span class="token number">1.9131586981302462e-08</span><span class="token punctuation">,</span> <span class="token number">2.111743526711507e-07</span><span class="token punctuation">,</span> <span class="token number">1.337269672774255e-08</span><span class="token punctuation">,</span> <span class="token number">2.0002481448955223e-08</span><span class="token punctuation">,</span> <span class="token number">1.0454256482717028e-07</span><span class="token punctuation">,</span> <span class="token number">2.8144228281234973e-08</span><span class="token punctuation">,</span> <span class="token number">2.1344791889532644e-07</span><span class="token punctuation">,</span> <span class="token number">2.1046110632028103e-08</span><span class="token punctuation">,</span> <span class="token number">1.9114453664315079e-07</span><span class="token punctuation">,</span> <span class="token number">3.957693550660224e-08</span><span class="token punctuation">,</span> <span class="token number">2.931631826186276e-08</span><span class="token punctuation">,</span> <span class="token number">1.105203111251285e-07</span><span class="token punctuation">,</span> <span class="token number">4.84007678380749e-08</span><span class="token punctuation">,</span> <span class="token number">5.583606110803885e-08</span><span class="token punctuation">,</span> <span class="token number">1.2130111315400427e-07</span><span class="token punctuation">,</span> <span class="token number">1.77621615193857e-08</span><span class="token punctuation">,</span> <span class="token number">2.5610853882085394e-08</span><span class="token punctuation">,</span> <span class="token number">1.203865309662433e-07</span><span class="token punctuation">,</span> <span class="token number">4.674859610531712e-09</span><span class="token punctuation">,</span> <span class="token number">1.5916098661250544e-08</span><span class="token punctuation">,</span> <span class="token number">3.147594185293201e-08</span><span class="token punctuation">,</span> <span class="token number">6.147686093527227e-08</span><span class="token punctuation">,</span> <span class="token number">2.204641802450169e-08</span><span class="token punctuation">,</span> <span class="token number">3.257763410147163e-07</span><span class="token punctuation">,</span> <span class="token number">1.198914532096751e-07</span><span class="token punctuation">,</span> <span class="token number">2.3818989802748547e-07</span><span class="token punctuation">,</span> <span class="token number">1.4909986134625797e-08</span><span class="token punctuation">,</span> <span class="token number">5.10168831624469e-08</span><span class="token punctuation">,</span> <span class="token number">5.5142201915714395e-08</span><span class="token punctuation">,</span> <span class="token number">2.288550327023131e-08</span><span class="token punctuation">,</span> <span class="token number">5.714110073995471e-08</span><span class="token punctuation">,</span> <span class="token number">5.185095801607531e-07</span><span class="token punctuation">,</span> <span class="token number">4.977285783525076e-08</span><span class="token punctuation">,</span> <span class="token number">1.1049896109227575e-08</span><span class="token punctuation">,</span> <span class="token number">1.264099296349741e-07</span><span class="token punctuation">,</span> <span class="token number">8.174881571676451e-08</span><span class="token punctuation">]</span><span class="token punctuation">]</span><span class="token punctuation">}</span>
</code></pre>
<h2>使用推理服务部署MLflow模型</h2>
<p>此示例将引导您了解如何利用KServe推理服务CRD部署mlflow模型，以及如何使用V2数据平面发送推理请求。</p>
<h3>训练</h3>
<p>第一步是通过调用mlflow log_model API来训练样本sklearn模型并保存为mlflow模型格式。</p>
<pre><code class="prism language-python"><span class="token comment"># Original source code and more details can be found in:</span>
<span class="token comment"># https://www.mlflow.org/docs/latest/tutorials-and-examples/tutorial.html</span>
<span class="token comment"># The data set used in this example is from</span>
<span class="token comment"># http://archive.ics.uci.edu/ml/datasets/Wine+Quality</span>
<span class="token comment"># P. Cortez, A. Cerdeira, F. Almeida, T. Matos and J. Reis.</span>
<span class="token comment"># Modeling wine preferences by data mining from physicochemical properties.</span>
<span class="token comment"># In Decision Support Systems, Elsevier, 47(4):547-553, 2009.</span>
<span class="token keyword">import</span> warnings
<span class="token keyword">import</span> sys
<span class="token keyword">import</span> pandas <span class="token keyword">as</span> pd
<span class="token keyword">import</span> numpy <span class="token keyword">as</span> np
<span class="token keyword">from</span> sklearn<span class="token punctuation">.</span>metrics <span class="token keyword">import</span> mean_squared_error<span class="token punctuation">,</span> mean_absolute_error<span class="token punctuation">,</span> r2_score
<span class="token keyword">from</span> sklearn<span class="token punctuation">.</span>model_selection <span class="token keyword">import</span> train_test_split
<span class="token keyword">from</span> sklearn<span class="token punctuation">.</span>linear_model <span class="token keyword">import</span> ElasticNet
<span class="token keyword">from</span> urllib<span class="token punctuation">.</span>parse <span class="token keyword">import</span> urlparse
<span class="token keyword">import</span> mlflow
<span class="token keyword">import</span> mlflow<span class="token punctuation">.</span>sklearn
<span class="token keyword">from</span> mlflow<span class="token punctuation">.</span>models<span class="token punctuation">.</span>signature <span class="token keyword">import</span> infer_signature
<span class="token keyword">import</span> logging
logging<span class="token punctuation">.</span>basicConfig<span class="token punctuation">(</span>level<span class="token operator">=</span>logging<span class="token punctuation">.</span>WARN<span class="token punctuation">)</span>
logger <span class="token operator">=</span> logging<span class="token punctuation">.</span>getLogger<span class="token punctuation">(</span>__name__<span class="token punctuation">)</span>
<span class="token keyword">def</span> <span class="token function">eval_metrics</span><span class="token punctuation">(</span>actual<span class="token punctuation">,</span> pred<span class="token punctuation">)</span><span class="token punctuation">:</span>
rmse <span class="token operator">=</span> np<span class="token punctuation">.</span>sqrt<span class="token punctuation">(</span>mean_squared_error<span class="token punctuation">(</span>actual<span class="token punctuation">,</span> pred<span class="token punctuation">)</span><span class="token punctuation">)</span>
mae <span class="token operator">=</span> mean_absolute_error<span class="token punctuation">(</span>actual<span class="token punctuation">,</span> pred<span class="token punctuation">)</span>
r2 <span class="token operator">=</span> r2_score<span class="token punctuation">(</span>actual<span class="token punctuation">,</span> pred<span class="token punctuation">)</span>
<span class="token keyword">return</span> rmse<span class="token punctuation">,</span> mae<span class="token punctuation">,</span> r2
<span class="token keyword">if</span> __name__ <span class="token operator">==</span> <span class="token string">"__main__"</span><span class="token punctuation">:</span>
warnings<span class="token punctuation">.</span>filterwarnings<span class="token punctuation">(</span><span class="token string">"ignore"</span><span class="token punctuation">)</span>
np<span class="token punctuation">.</span>random<span class="token punctuation">.</span>seed<span class="token punctuation">(</span><span class="token number">40</span><span class="token punctuation">)</span>
<span class="token comment"># Read the wine-quality csv file from the URL</span>
csv_url <span class="token operator">=</span> <span class="token punctuation">(</span>
<span class="token string">"http://archive.ics.uci.edu/ml"</span>
<span class="token string">"/machine-learning-databases/wine-quality/winequality-red.csv"</span>
<span class="token punctuation">)</span>
<span class="token keyword">try</span><span class="token punctuation">:</span>
data <span class="token operator">=</span> pd<span class="token punctuation">.</span>read_csv<span class="token punctuation">(</span>csv_url<span class="token punctuation">,</span> sep<span class="token operator">=</span><span class="token string">";"</span><span class="token punctuation">)</span>
<span class="token keyword">except</span> Exception <span class="token keyword">as</span> e<span class="token punctuation">:</span>
logger<span class="token punctuation">.</span>exception<span class="token punctuation">(</span>
<span class="token string">"Unable to download training & test CSV, "</span>
<span class="token string">"check your internet connection. Error: %s"</span><span class="token punctuation">,</span>
e<span class="token punctuation">,</span>
<span class="token punctuation">)</span>
<span class="token comment"># Split the data into training and test sets. (0.75, 0.25) split.</span>
train<span class="token punctuation">,</span> test <span class="token operator">=</span> train_test_split<span class="token punctuation">(</span>data<span class="token punctuation">)</span>
<span class="token comment"># The predicted column is "quality" which is a scalar from [3, 9]</span>
train_x <span class="token operator">=</span> train<span class="token punctuation">.</span>drop<span class="token punctuation">(</span><span class="token punctuation">[</span><span class="token string">"quality"</span><span class="token punctuation">]</span><span class="token punctuation">,</span> axis<span class="token operator">=</span><span class="token number">1</span><span class="token punctuation">)</span>
test_x <span class="token operator">=</span> test<span class="token punctuation">.</span>drop<span class="token punctuation">(</span><span class="token punctuation">[</span><span class="token string">"quality"</span><span class="token punctuation">]</span><span class="token punctuation">,</span> axis<span class="token operator">=</span><span class="token number">1</span><span class="token punctuation">)</span>
train_y <span class="token operator">=</span> train<span class="token punctuation">[</span><span class="token punctuation">[</span><span class="token string">"quality"</span><span class="token punctuation">]</span><span class="token punctuation">]</span>
test_y <span class="token operator">=</span> test<span class="token punctuation">[</span><span class="token punctuation">[</span><span class="token string">"quality"</span><span class="token punctuation">]</span><span class="token punctuation">]</span>
alpha <span class="token operator">=</span> <span class="token builtin">float</span><span class="token punctuation">(</span>sys<span class="token punctuation">.</span>argv<span class="token punctuation">[</span><span class="token number">1</span><span class="token punctuation">]</span><span class="token punctuation">)</span> <span class="token keyword">if</span> <span class="token builtin">len</span><span class="token punctuation">(</span>sys<span class="token punctuation">.</span>argv<span class="token punctuation">)</span> <span class="token operator">></span> <span class="token number">1</span> <span class="token keyword">else</span> <span class="token number">0.5</span>
l1_ratio <span class="token operator">=</span> <span class="token builtin">float</span><span class="token punctuation">(</span>sys<span class="token punctuation">.</span>argv<span class="token punctuation">[</span><span class="token number">2</span><span class="token punctuation">]</span><span class="token punctuation">)</span> <span class="token keyword">if</span> <span class="token builtin">len</span><span class="token punctuation">(</span>sys<span class="token punctuation">.</span>argv<span class="token punctuation">)</span> <span class="token operator">></span> <span class="token number">2</span> <span class="token keyword">else</span> <span class="token number">0.5</span>
<span class="token keyword">with</span> mlflow<span class="token punctuation">.</span>start_run<span class="token punctuation">(</span><span class="token punctuation">)</span><span class="token punctuation">:</span>
lr <span class="token operator">=</span> ElasticNet<span class="token punctuation">(</span>alpha<span class="token operator">=</span>alpha<span class="token punctuation">,</span> l1_ratio<span class="token operator">=</span>l1_ratio<span class="token punctuation">,</span> random_state<span class="token operator">=</span><span class="token number">42</span><span class="token punctuation">)</span>
lr<span class="token punctuation">.</span>fit<span class="token punctuation">(</span>train_x<span class="token punctuation">,</span> train_y<span class="token punctuation">)</span>
predicted_qualities <span class="token operator">=</span> lr<span class="token punctuation">.</span>predict<span class="token punctuation">(</span>test_x<span class="token punctuation">)</span>
<span class="token punctuation">(</span>rmse<span class="token punctuation">,</span> mae<span class="token punctuation">,</span> r2<span class="token punctuation">)</span> <span class="token operator">=</span> eval_metrics<span class="token punctuation">(</span>test_y<span class="token punctuation">,</span> predicted_qualities<span class="token punctuation">)</span>
<span class="token keyword">print</span><span class="token punctuation">(</span><span class="token string">"Elasticnet model (alpha=%f, l1_ratio=%f):"</span> <span class="token operator">%</span> <span class="token punctuation">(</span>alpha<span class="token punctuation">,</span> l1_ratio<span class="token punctuation">)</span><span class="token punctuation">)</span>
<span class="token keyword">print</span><span class="token punctuation">(</span><span class="token string">" RMSE: %s"</span> <span class="token operator">%</span> rmse<span class="token punctuation">)</span>
<span class="token keyword">print</span><span class="token punctuation">(</span><span class="token string">" MAE: %s"</span> <span class="token operator">%</span> mae<span class="token punctuation">)</span>
<span class="token keyword">print</span><span class="token punctuation">(</span><span class="token string">" R2: %s"</span> <span class="token operator">%</span> r2<span class="token punctuation">)</span>
mlflow<span class="token punctuation">.</span>log_param<span class="token punctuation">(</span><span class="token string">"alpha"</span><span class="token punctuation">,</span> alpha<span class="token punctuation">)</span>
mlflow<span class="token punctuation">.</span>log_param<span class="token punctuation">(</span><span class="token string">"l1_ratio"</span><span class="token punctuation">,</span> l1_ratio<span class="token punctuation">)</span>
mlflow<span class="token punctuation">.</span>log_metric<span class="token punctuation">(</span><span class="token string">"rmse"</span><span class="token punctuation">,</span> rmse<span class="token punctuation">)</span>
mlflow<span class="token punctuation">.</span>log_metric<span class="token punctuation">(</span><span class="token string">"r2"</span><span class="token punctuation">,</span> r2<span class="token punctuation">)</span>
mlflow<span class="token punctuation">.</span>log_metric<span class="token punctuation">(</span><span class="token string">"mae"</span><span class="token punctuation">,</span> mae<span class="token punctuation">)</span>
tracking_url_type_store <span class="token operator">=</span> urlparse<span class="token punctuation">(</span>mlflow<span class="token punctuation">.</span>get_tracking_uri<span class="token punctuation">(</span><span class="token punctuation">)</span><span class="token punctuation">)</span><span class="token punctuation">.</span>scheme
model_signature <span class="token operator">=</span> infer_signature<span class="token punctuation">(</span>train_x<span class="token punctuation">,</span> train_y<span class="token punctuation">)</span>
<span class="token comment"># Model registry does not work with file store</span>
<span class="token keyword">if</span> tracking_url_type_store <span class="token operator">!=</span> <span class="token string">"file"</span><span class="token punctuation">:</span>
<span class="token comment"># Register the model</span>
<span class="token comment"># There are other ways to use the Model Registry,</span>
<span class="token comment"># which depends on the use case,</span>
<span class="token comment"># please refer to the doc for more information:</span>
<span class="token comment"># https://mlflow.org/docs/latest/model-registry.html#api-workflow</span>
mlflow<span class="token punctuation">.</span>sklearn<span class="token punctuation">.</span>log_model<span class="token punctuation">(</span>
lr<span class="token punctuation">,</span>
<span class="token string">"model"</span><span class="token punctuation">,</span>
registered_model_name<span class="token operator">=</span><span class="token string">"ElasticnetWineModel"</span><span class="token punctuation">,</span>
signature<span class="token operator">=</span>model_signature<span class="token punctuation">,</span>
<span class="token punctuation">)</span>
<span class="token keyword">else</span><span class="token punctuation">:</span>
mlflow<span class="token punctuation">.</span>sklearn<span class="token punctuation">.</span>log_model<span class="token punctuation">(</span>lr<span class="token punctuation">,</span> <span class="token string">"model"</span><span class="token punctuation">,</span> signature<span class="token operator">=</span>model_signature<span class="token punctuation">)</span>
</code></pre>
<p>训练脚本还将利用MLflow模型格式对我们训练的模型进行序列化。</p>
<pre><code>model/
├── MLmodel
├── model.pkl
├── conda.yaml
└── requirements.txt
</code></pre>
<h3>本地测试</h3>
<p>一旦你得到了你的模型序列化model.pkl，我们就可以使用MLServer来启动一个本地服务器。有关MLServer的更多详细信息，请随时查看MLflow示例文档。</p>
<blockquote>
<p>笔记<br> 此步骤是可选的，仅用于测试，可以直接使用InfenceService进行部署。</p>
</blockquote>
<h4>先决条件</h4>
<p>首先，要在本地使用MLServer，首先需要在本地环境以及MLflow运行时中安装MLServer包。</p>
<pre><code class="prism language-powershell">pip install mlserver mlserver-mlflow
</code></pre>
<h4>模型设置</h4>
<p>下一步将提供一些模型设置，以便MLServer知道：</p>
<ul>
<li>为模型提供服务的推理运行时（即mlserver_mlflow.MLflowRuntime）</li>
<li>型号的名称和版本</li>
</ul>
<p>这些可以通过环境变量或创建本地model-settings.json文件来指定：</p>
<pre><code class="prism language-python"><span class="token punctuation">{</span>
<span class="token string">"name"</span><span class="token punctuation">:</span> <span class="token string">"mlflow-wine-classifier"</span><span class="token punctuation">,</span>
<span class="token string">"version"</span><span class="token punctuation">:</span> <span class="token string">"v1.0.0"</span><span class="token punctuation">,</span>
<span class="token string">"implementation"</span><span class="token punctuation">:</span> <span class="token string">"mlserver_mlflow.MLflowRuntime"</span>
<span class="token punctuation">}</span>
</code></pre>
<h4>在本地启动模型服务器</h4>
<p>有了本地安装的mlserver软件包和本地model-settings.json文件，您现在应该可以启动我们的服务器了：</p>
<pre><code>mlserver start .
</code></pre>
<h3>使用推理服务进行部署</h3>
<p>当您使用InferenceService部署模型时，KServe会注入合理的默认值，这样它就可以在没有任何进一步配置的情况下开箱即用。但是，您仍然可以通过提供类似于本地文件的model-settings.json文件来覆盖这些默认值。您甚至可以提供一组model-settings.json文件来加载多个模型。</p>
<p>为了使用v2协议对部署的模型进行推理，您将protocolVersion字段设置为v2，在本例中，您的模型工件已经上传到“GCS模型存储库”，可以作为gs://kfserving examples/models/mlflow/wine访问。<br> <code>新框架</code></p>
<pre><code class="prism language-yaml"><span class="token key atrule">apiVersion</span><span class="token punctuation">:</span> <span class="token string">"serving.kserve.io/v1beta1"</span>
<span class="token key atrule">kind</span><span class="token punctuation">:</span> <span class="token string">"InferenceService"</span>
<span class="token key atrule">metadata</span><span class="token punctuation">:</span>
<span class="token key atrule">name</span><span class="token punctuation">:</span> <span class="token string">"mlflow-v2-wine-classifier"</span>
<span class="token key atrule">spec</span><span class="token punctuation">:</span>
<span class="token key atrule">predictor</span><span class="token punctuation">:</span>
<span class="token key atrule">model</span><span class="token punctuation">:</span>
<span class="token key atrule">modelFormat</span><span class="token punctuation">:</span>
<span class="token key atrule">name</span><span class="token punctuation">:</span> mlflow
<span class="token key atrule">protocolVersion</span><span class="token punctuation">:</span> v2
<span class="token key atrule">storageUri</span><span class="token punctuation">:</span> <span class="token string">"gs://kfserving-examples/models/mlflow/wine"</span>
</code></pre>
<p><code>kubectl</code></p>
<pre><code>kubectl apply -f mlflow-new.yaml
</code></pre>
<h3>测试部署的模型</h3>
<p>现在，您可以通过发送示例请求来测试已部署的模型。</p>
<p>请注意，此请求需要遵循V2数据平面协议。您可以在下面看到一个负载示例：</p>
<pre><code class="prism language-python"><span class="token punctuation">{</span>
<span class="token string">"parameters"</span><span class="token punctuation">:</span> <span class="token punctuation">{</span>
<span class="token string">"content_type"</span><span class="token punctuation">:</span> <span class="token string">"pd"</span>
<span class="token punctuation">}</span><span class="token punctuation">,</span>
<span class="token string">"inputs"</span><span class="token punctuation">:</span> <span class="token punctuation">[</span>
<span class="token punctuation">{</span>
<span class="token string">"name"</span><span class="token punctuation">:</span> <span class="token string">"fixed acidity"</span><span class="token punctuation">,</span>
<span class="token string">"shape"</span><span class="token punctuation">:</span> <span class="token punctuation">[</span><span class="token number">1</span><span class="token punctuation">]</span><span class="token punctuation">,</span>
<span class="token string">"datatype"</span><span class="token punctuation">:</span> <span class="token string">"FP32"</span><span class="token punctuation">,</span>
<span class="token string">"data"</span><span class="token punctuation">:</span> <span class="token punctuation">[</span><span class="token number">7.4</span><span class="token punctuation">]</span>
<span class="token punctuation">}</span><span class="token punctuation">,</span>
<span class="token punctuation">{</span>
<span class="token string">"name"</span><span class="token punctuation">:</span> <span class="token string">"volatile acidity"</span><span class="token punctuation">,</span>
<span class="token string">"shape"</span><span class="token punctuation">:</span> <span class="token punctuation">[</span><span class="token number">1</span><span class="token punctuation">]</span><span class="token punctuation">,</span>
<span class="token string">"datatype"</span><span class="token punctuation">:</span> <span class="token string">"FP32"</span><span class="token punctuation">,</span>
<span class="token string">"data"</span><span class="token punctuation">:</span> <span class="token punctuation">[</span><span class="token number">0.7000</span><span class="token punctuation">]</span>
<span class="token punctuation">}</span><span class="token punctuation">,</span>
<span class="token punctuation">{</span>
<span class="token string">"name"</span><span class="token punctuation">:</span> <span class="token string">"citric acid"</span><span class="token punctuation">,</span>
<span class="token string">"shape"</span><span class="token punctuation">:</span> <span class="token punctuation">[</span><span class="token number">1</span><span class="token punctuation">]</span><span class="token punctuation">,</span>
<span class="token string">"datatype"</span><span class="token punctuation">:</span> <span class="token string">"FP32"</span><span class="token punctuation">,</span>
<span class="token string">"data"</span><span class="token punctuation">:</span> <span class="token punctuation">[</span><span class="token number">0</span><span class="token punctuation">]</span>
<span class="token punctuation">}</span><span class="token punctuation">,</span>
<span class="token punctuation">{</span>
<span class="token string">"name"</span><span class="token punctuation">:</span> <span class="token string">"residual sugar"</span><span class="token punctuation">,</span>
<span class="token string">"shape"</span><span class="token punctuation">:</span> <span class="token punctuation">[</span><span class="token number">1</span><span class="token punctuation">]</span><span class="token punctuation">,</span>
<span class="token string">"datatype"</span><span class="token punctuation">:</span> <span class="token string">"FP32"</span><span class="token punctuation">,</span>
<span class="token string">"data"</span><span class="token punctuation">:</span> <span class="token punctuation">[</span><span class="token number">1.9</span><span class="token punctuation">]</span>
<span class="token punctuation">}</span><span class="token punctuation">,</span>
<span class="token punctuation">{</span>
<span class="token string">"name"</span><span class="token punctuation">:</span> <span class="token string">"chlorides"</span><span class="token punctuation">,</span>
<span class="token string">"shape"</span><span class="token punctuation">:</span> <span class="token punctuation">[</span><span class="token number">1</span><span class="token punctuation">]</span><span class="token punctuation">,</span>
<span class="token string">"datatype"</span><span class="token punctuation">:</span> <span class="token string">"FP32"</span><span class="token punctuation">,</span>
<span class="token string">"data"</span><span class="token punctuation">:</span> <span class="token punctuation">[</span><span class="token number">0.076</span><span class="token punctuation">]</span>
<span class="token punctuation">}</span><span class="token punctuation">,</span>
<span class="token punctuation">{</span>
<span class="token string">"name"</span><span class="token punctuation">:</span> <span class="token string">"free sulfur dioxide"</span><span class="token punctuation">,</span>
<span class="token string">"shape"</span><span class="token punctuation">:</span> <span class="token punctuation">[</span><span class="token number">1</span><span class="token punctuation">]</span><span class="token punctuation">,</span>
<span class="token string">"datatype"</span><span class="token punctuation">:</span> <span class="token string">"FP32"</span><span class="token punctuation">,</span>
<span class="token string">"data"</span><span class="token punctuation">:</span> <span class="token punctuation">[</span><span class="token number">11</span><span class="token punctuation">]</span>
<span class="token punctuation">}</span><span class="token punctuation">,</span>
<span class="token punctuation">{</span>
<span class="token string">"name"</span><span class="token punctuation">:</span> <span class="token string">"total sulfur dioxide"</span><span class="token punctuation">,</span>
<span class="token string">"shape"</span><span class="token punctuation">:</span> <span class="token punctuation">[</span><span class="token number">1</span><span class="token punctuation">]</span><span class="token punctuation">,</span>
<span class="token string">"datatype"</span><span class="token punctuation">:</span> <span class="token string">"FP32"</span><span class="token punctuation">,</span>
<span class="token string">"data"</span><span class="token punctuation">:</span> <span class="token punctuation">[</span><span class="token number">34</span><span class="token punctuation">]</span>
<span class="token punctuation">}</span><span class="token punctuation">,</span>
<span class="token punctuation">{</span>
<span class="token string">"name"</span><span class="token punctuation">:</span> <span class="token string">"density"</span><span class="token punctuation">,</span>
<span class="token string">"shape"</span><span class="token punctuation">:</span> <span class="token punctuation">[</span><span class="token number">1</span><span class="token punctuation">]</span><span class="token punctuation">,</span>
<span class="token string">"datatype"</span><span class="token punctuation">:</span> <span class="token string">"FP32"</span><span class="token punctuation">,</span>
<span class="token string">"data"</span><span class="token punctuation">:</span> <span class="token punctuation">[</span><span class="token number">0.9978</span><span class="token punctuation">]</span>
<span class="token punctuation">}</span><span class="token punctuation">,</span>
<span class="token punctuation">{</span>
<span class="token string">"name"</span><span class="token punctuation">:</span> <span class="token string">"pH"</span><span class="token punctuation">,</span>
<span class="token string">"shape"</span><span class="token punctuation">:</span> <span class="token punctuation">[</span><span class="token number">1</span><span class="token punctuation">]</span><span class="token punctuation">,</span>
<span class="token string">"datatype"</span><span class="token punctuation">:</span> <span class="token string">"FP32"</span><span class="token punctuation">,</span>
<span class="token string">"data"</span><span class="token punctuation">:</span> <span class="token punctuation">[</span><span class="token number">3.51</span><span class="token punctuation">]</span>
<span class="token punctuation">}</span><span class="token punctuation">,</span>
<span class="token punctuation">{</span>
<span class="token string">"name"</span><span class="token punctuation">:</span> <span class="token string">"sulphates"</span><span class="token punctuation">,</span>
<span class="token string">"shape"</span><span class="token punctuation">:</span> <span class="token punctuation">[</span><span class="token number">1</span><span class="token punctuation">]</span><span class="token punctuation">,</span>
<span class="token string">"datatype"</span><span class="token punctuation">:</span> <span class="token string">"FP32"</span><span class="token punctuation">,</span>
<span class="token string">"data"</span><span class="token punctuation">:</span> <span class="token punctuation">[</span><span class="token number">0.56</span><span class="token punctuation">]</span>
<span class="token punctuation">}</span><span class="token punctuation">,</span>
<span class="token punctuation">{</span>
<span class="token string">"name"</span><span class="token punctuation">:</span> <span class="token string">"alcohol"</span><span class="token punctuation">,</span>
<span class="token string">"shape"</span><span class="token punctuation">:</span> <span class="token punctuation">[</span><span class="token number">1</span><span class="token punctuation">]</span><span class="token punctuation">,</span>
<span class="token string">"datatype"</span><span class="token punctuation">:</span> <span class="token string">"FP32"</span><span class="token punctuation">,</span>
<span class="token string">"data"</span><span class="token punctuation">:</span> <span class="token punctuation">[</span><span class="token number">9.4</span><span class="token punctuation">]</span>
<span class="token punctuation">}</span>
<span class="token punctuation">]</span>
<span class="token punctuation">}</span>
</code></pre>
<p>现在，假设您的入口可以在${INGRESS_HOST}: ${INGRESS_PORT}访问，或者您可以按照此说明查找入口IP和端口。</p>
<p>您可以使用curl将推理请求发送为：</p>
<pre><code class="prism language-powershell">SERVICE_HOSTNAME=$<span class="token punctuation">(</span>kubectl get inferenceservice mlflow-v2-wine-classifier <span class="token operator">-</span>o jsonpath=<span class="token string">'{.status.url}'</span> <span class="token punctuation">|</span> cut <span class="token operator">-</span>d <span class="token string">"/"</span> <span class="token operator">-</span>f 3<span class="token punctuation">)</span>
curl <span class="token operator">-</span>v \
<span class="token operator">-</span>H <span class="token string">"Host: ${SERVICE_HOSTNAME}"</span> \
<span class="token operator">-</span>H <span class="token string">"Content-Type: application/json"</span> \
<span class="token operator">-</span>d @<span class="token punctuation">.</span><span class="token operator">/</span>mlflow-input<span class="token punctuation">.</span>json \
http:<span class="token operator">/</span><span class="token operator">/</span>$<span class="token punctuation">{</span>INGRESS_HOST<span class="token punctuation">}</span>:$<span class="token punctuation">{</span>INGRESS_PORT<span class="token punctuation">}</span><span class="token operator">/</span>v2/models/mlflow-v2-wine-classifier/infer
</code></pre>
<p><code>期望输出</code></p>
<pre><code class="prism language-powershell"><span class="token punctuation">{</span>
<span class="token string">"model_name"</span>:<span class="token string">"mlflow-v2-wine-classifier"</span><span class="token punctuation">,</span>
<span class="token string">"model_version"</span>:null<span class="token punctuation">,</span>
<span class="token string">"id"</span>:<span class="token string">"699cf11c-e843-444e-9dc3-000d991052cc"</span><span class="token punctuation">,</span>
<span class="token string">"parameters"</span>:null<span class="token punctuation">,</span>
<span class="token string">"outputs"</span>:<span class="token punctuation">[</span>
<span class="token punctuation">{</span>
<span class="token string">"name"</span>:<span class="token string">"predict"</span><span class="token punctuation">,</span>
<span class="token string">"shape"</span>:<span class="token punctuation">[</span>1<span class="token punctuation">]</span><span class="token punctuation">,</span>
<span class="token string">"datatype"</span>:<span class="token string">"FP64"</span><span class="token punctuation">,</span>
<span class="token string">"parameters"</span>:null<span class="token punctuation">,</span>
<span class="token string">"data"</span>:<span class="token punctuation">[</span>5<span class="token punctuation">.</span>576883936610762<span class="token punctuation">]</span>
<span class="token punctuation">}</span>
<span class="token punctuation">]</span>
<span class="token punctuation">}</span>
</code></pre>
</div>
</div>
</div>
</div>
</div>
<!--PC和WAP自适应版-->
<div id="SOHUCS" sid="1705436075637420032"></div>
<script type="text/javascript" src="/views/front/js/chanyan.js"></script>
<!-- 文章页-底部 动态广告位 -->
<div class="youdao-fixed-ad" id="detail_ad_bottom"></div>
</div>
<div class="col-md-3">
<div class="row" id="ad">
<!-- 文章页-右侧1 动态广告位 -->
<div id="right-1" class="col-lg-12 col-md-12 col-sm-4 col-xs-4 ad">
<div class="youdao-fixed-ad" id="detail_ad_1"> </div>
</div>
<!-- 文章页-右侧2 动态广告位 -->
<div id="right-2" class="col-lg-12 col-md-12 col-sm-4 col-xs-4 ad">
<div class="youdao-fixed-ad" id="detail_ad_2"></div>
</div>
<!-- 文章页-右侧3 动态广告位 -->
<div id="right-3" class="col-lg-12 col-md-12 col-sm-4 col-xs-4 ad">
<div class="youdao-fixed-ad" id="detail_ad_3"></div>
</div>
</div>
</div>
</div>
</div>
</div>
<div class="container">
<h4 class="pt20 mb15 mt0 border-top">你可能感兴趣的:(Kserve,Kserve)</h4>
<div id="paradigm-article-related">
<div class="recommend-post mb30">
<ul class="widget-links">
<li><a href="/article/1721063468603355136.htm"
title="从minio下载zip文件报错，无响应" target="_blank">从minio下载zip文件报错，无响应</a>
<span class="text-muted">Rory602</span>
<a class="tag" taget="_blank" href="/search/k8s/1.htm">k8s</a><a class="tag" taget="_blank" href="/search/minio/1.htm">minio</a><a class="tag" taget="_blank" href="/search/kubernetes/1.htm">kubernetes</a>
<div>1.问题描述：Traceback(mostrecentcalllast):File"/storage-initializer/scripts/initializer-entrypoint",line14,inkserve.Storage.download(src_uri,dest_path)File"/usr/local/lib/python3.7/site-packages/kserve/sto</div>
</li>
<li><a href="/article/1705436201810472960.htm"
title="KServe Documentation Website" target="_blank">KServe Documentation Website</a>
<span class="text-muted">蓝净云</span>
<a class="tag" taget="_blank" href="/search/Kserve/1.htm">Kserve</a><a class="tag" taget="_blank" href="/search/%E6%B1%87%E6%80%BB/1.htm">汇总</a><a class="tag" taget="_blank" href="/search/Kserve/1.htm">Kserve</a>
<div>KServeDocumentation-UserGuide翻译Concepts概念ModelServingRuntimes模型服务运行时、、SupportedModelFrameworks/Formats支持的模型框架/格式、、Multi-FrameworkServingRuntimes多框架服务运行时、、Howtowriteacustompredictor如何编写自定义预测器MultiModel</div>
</li>
<li><a href="/article/1705436202284429312.htm"
title="Multi-Framework Serving Runtimes-How to write a custom predictor-KServe" target="_blank">Multi-Framework Serving Runtimes-How to write a custom predictor-KServe</a>
<span class="text-muted">蓝净云</span>
<a class="tag" taget="_blank" href="/search/Kserve/1.htm">Kserve</a><a class="tag" taget="_blank" href="/search/Kserve/1.htm">Kserve</a>
<div>Multi-FrameworkServingRuntimes-Howtowriteacustompredictor用TorchScript模型预测Triton推理服务设置导出为Torchscript模型将您在GCS上训练的模型存储在模型库中使用HTTP端点推断创建推理服务使用curl运行预测运行性能测试gRPC端点推断创建推理服务使用grpcurl运行预测将Transformer添加到推理服务实现</div>
</li>
<li><a href="/article/1705436075637420032.htm"
title="Supported Model Frameworks/Formats - KServe" target="_blank">Supported Model Frameworks/Formats - KServe</a>
<span class="text-muted">蓝净云</span>
<a class="tag" taget="_blank" href="/search/Kserve/1.htm">Kserve</a><a class="tag" taget="_blank" href="/search/Kserve/1.htm">Kserve</a>
<div>支持的模型框架/格式模型服务运行时间使用推理服务部署Tensorflow模型创建HTTP推理服务运行预测金丝雀Rollout创建gRPC推理服务运行预测使用TorchServe推理服务部署PyTorch模型使用模型档案文件和配置创建模型存储使用V1REST协议部署PyTorch模型创建TorchServe推理服务模型推理模型说明使用V2REST协议部署PyTorch模型创建推理服务模型推理模型说明</div>
</li>
<li><a href="/article/1691794163663712256.htm"
title="如何基于 ACK Serverless 快速部署 AI 推理服务" target="_blank">如何基于 ACK Serverless 快速部署 AI 推理服务</a>
<span class="text-muted"></span>
<div>作者：元毅随着AI浪潮的到来，各种AI应用层出不穷，众所周知AI应用对GPU资源强烈依赖，但GPU很昂贵，如何降低GPU资源使用成本成为用户首要问题。而AI与Serverless技术结合，完全可以达到按需使用资源，降低资源成本的目的。那么在云原生场景下，是否有这样开箱即用、标准、开放的方案呢？答案是有。我们在ACKServerless中提供Knative+KServe的方案，可以帮助用户快速部署A</div>
</li>
<li><a href="/article/1691415140798377984.htm"
title="如何基于 ACK Serverless 快速部署 AI 推理服务" target="_blank">如何基于 ACK Serverless 快速部署 AI 推理服务</a>
<span class="text-muted"></span>
<div>作者：元毅随着AI浪潮的到来，各种AI应用层出不穷，众所周知AI应用对GPU资源强烈依赖，但GPU很昂贵，如何降低GPU资源使用成本成为用户首要问题。而AI与Serverless技术结合，完全可以达到按需使用资源，降低资源成本的目的。那么在云原生场景下，是否有这样开箱即用、标准、开放的方案呢？答案是有。我们在ACKServerless中提供Knative+KServe的方案，可以帮助用户快速部署A</div>
</li>
<li><a href="/article/1691415130732048384.htm"
title="【AI 充电】KServe + Fluid 加速大模型推理" target="_blank">【AI 充电】KServe + Fluid 加速大模型推理</a>
<span class="text-muted"></span>
<a class="tag" taget="_blank" href="/search/%E9%98%BF%E9%87%8C%E4%BA%91%E4%BA%91%E5%8E%9F%E7%94%9F%E5%AE%B9%E5%99%A8/1.htm">阿里云云原生容器</a>
<div>作者：黄驰琳、露营、车漾背景KServe是Kubernetes上的标准模型推理平台，专为高度可扩展的场景而构建，支持现代Serverless推理工作负载，用于在任意框架上提供机器学习（ML）模型服务。它提供高性能、高度抽象的接口，以支持常见的ML框架（如Tensorflow、XGBoost、Scikit-Learn、PyTorch和ONNX）来解决生产模型服务场景。此外，KServe封装了自动缩放</div>
</li>
<li><a href="/article/1691307081111515136.htm"
title="如何基于 ACK Serverless 快速部署 AI 推理服务" target="_blank">如何基于 ACK Serverless 快速部署 AI 推理服务</a>
<span class="text-muted">阿里云云原生</span>
<a class="tag" taget="_blank" href="/search/serverless/1.htm">serverless</a><a class="tag" taget="_blank" href="/search/%E4%BA%BA%E5%B7%A5%E6%99%BA%E8%83%BD/1.htm">人工智能</a><a class="tag" taget="_blank" href="/search/%E4%BA%91%E5%8E%9F%E7%94%9F/1.htm">云原生</a>
<div>作者：元毅随着AI浪潮的到来，各种AI应用层出不穷，众所周知AI应用对GPU资源强烈依赖，但GPU很昂贵，如何降低GPU资源使用成本成为用户首要问题。而AI与Serverless技术结合，完全可以达到按需使用资源，降低资源成本的目的。那么在云原生场景下，是否有这样开箱即用、标准、开放的方案呢？答案是有。我们在ACKServerless中提供Knative+KServe的方案，可以帮助用户快速部署A</div>
</li>
<li><a href="/article/1687330906353774592.htm"
title="基于Kubernetes环境的高扩展机器学习部署利器——KServe" target="_blank">基于Kubernetes环境的高扩展机器学习部署利器——KServe</a>
<span class="text-muted">冲浪中台</span>
<a class="tag" taget="_blank" href="/search/kubernetes/1.htm">kubernetes</a><a class="tag" taget="_blank" href="/search/%E6%9C%BA%E5%99%A8%E5%AD%A6%E4%B9%A0/1.htm">机器学习</a><a class="tag" taget="_blank" href="/search/%E5%AE%B9%E5%99%A8/1.htm">容器</a>
<div>随着ChatGPT的发布，人们越来越难以回避利用机器学习的相关技术。从消息应用程序上的文本预测到智能门铃上的面部识别，机器学习（ML）几乎可以在我们今天使用的每一项技术中找到。如何将机器学习技术交付给消费者是企业在开发过程中必须解决的众多挑战之一。机器学习产品的部署策略对产品的最终用户有重大影响。这可能意味着，iPhone上的Siri和网络浏览器中的ChatGPT之间将存在重大差异。除了ChatG</div>
</li>
<li><a href="/article/1681136046303162368.htm"
title="【AI 充电】KServe + Fluid 加速大模型推理" target="_blank">【AI 充电】KServe + Fluid 加速大模型推理</a>
<span class="text-muted">阿里云云原生</span>
<a class="tag" taget="_blank" href="/search/%E4%BA%BA%E5%B7%A5%E6%99%BA%E8%83%BD/1.htm">人工智能</a><a class="tag" taget="_blank" href="/search/%E9%98%BF%E9%87%8C%E4%BA%91/1.htm">阿里云</a><a class="tag" taget="_blank" href="/search/%E4%BA%91%E5%8E%9F%E7%94%9F/1.htm">云原生</a>
<div>作者：黄驰琳、露营、车漾背景KServe是Kubernetes上的标准模型推理平台，专为高度可扩展的场景而构建，支持现代Serverless推理工作负载，用于在任意框架上提供机器学习（ML）模型服务。它提供高性能、高度抽象的接口，以支持常见的ML框架（如Tensorflow、XGBoost、Scikit-Learn、PyTorch和ONNX）来解决生产模型服务场景。此外，KServe封装了自动缩放</div>
</li>
<li><a href="/article/126.htm"
title="插入表主键冲突做更新" target="_blank">插入表主键冲突做更新</a>
<span class="text-muted">a-john</span>
<div>有以下场景：
用户下了一个订单，订单内的内容较多，且来自多表，首次下单的时候，内容可能会不全（部分内容不是必须，出现有些表根本就没有没有该订单的值）。在以后更改订单时，有些内容会更改，有些内容会新增。
问题：
如果在sql语句中执行update操作，在没有数据的表中会出错。如果在逻辑代码中先做查询，查询结果有做更新，没有做插入，这样会将代码复杂化。
解决：
mysql中提供了一个sql语</div>
</li>
<li><a href="/article/253.htm"
title="Android xml资源文件中@、@android:type、@*、？、@+含义和区别" target="_blank">Android xml资源文件中@、@android:type、@*、？、@+含义和区别</a>
<span class="text-muted">Cb123456</span>
<a class="tag" taget="_blank" href="/search/%40%2B/1.htm">@+</a><a class="tag" taget="_blank" href="/search/%40/1.htm">@</a><a class="tag" taget="_blank" href="/search/%3F/1.htm">?</a><a class="tag" taget="_blank" href="/search/%40*/1.htm">@*</a>
<div>一.@代表引用资源
1.引用自定义资源。格式：@[package:]type/name
android：text="@string/hello"
2.引用系统资源。格式：@android:type/name
android:textColor="@android:color/opaque_red"</div>
</li>
<li><a href="/article/380.htm"
title="数据结构的基本介绍" target="_blank">数据结构的基本介绍</a>
<span class="text-muted">天子之骄</span>
<a class="tag" taget="_blank" href="/search/%E6%95%B0%E6%8D%AE%E7%BB%93%E6%9E%84/1.htm">数据结构</a><a class="tag" taget="_blank" href="/search/%E6%95%A3%E5%88%97%E8%A1%A8/1.htm">散列表</a><a class="tag" taget="_blank" href="/search/%E6%A0%91%E3%80%81%E5%9B%BE/1.htm">树、图</a><a class="tag" taget="_blank" href="/search/%E7%BA%BF%E6%80%A7%E7%BB%93%E6%9E%84/1.htm">线性结构</a><a class="tag" taget="_blank" href="/search/%E4%BB%B7%E6%A0%BC%E6%A0%87%E7%AD%BE/1.htm">价格标签</a>
<div>数据结构的基本介绍
数据结构就是数据的组织形式，用一种提前设计好的框架去存取数据，以便更方便，高效的对数据进行增删查改。正确选择合适的数据结构，对软件程序的高效执行的影响作用不亚于算法的设计。此外，在计算机系统中数据结构的作用也是非同小可。例如常常在编程语言中听到的栈，堆等，就是经典的数据结构。
经典的数据结构大致如下：
一：线性数据结构
(1)：列表
a</div>
</li>
<li><a href="/article/507.htm"
title="通过二维码开放平台的API快速生成二维码" target="_blank">通过二维码开放平台的API快速生成二维码</a>
<span class="text-muted">一炮送你回车库</span>
<a class="tag" taget="_blank" href="/search/api/1.htm">api</a>
<div> 现在很多网站都有通过扫二维码用手机连接的功能，联图网(http://www.liantu.com/pingtai/)的二维码开放平台开放了一个生成二维码图片的Api,挺方便使用的。闲着无聊，写了个前台快速生成二维码的方法。
html代码如下:(二维码将生成在这div下)
? 1
&nbs</div>
</li>
<li><a href="/article/634.htm"
title="ImageIO读取一张图片改变大小" target="_blank">ImageIO读取一张图片改变大小</a>
<span class="text-muted">3213213333332132</span>
<a class="tag" taget="_blank" href="/search/java/1.htm">java</a><a class="tag" taget="_blank" href="/search/IO/1.htm">IO</a><a class="tag" taget="_blank" href="/search/image/1.htm">image</a><a class="tag" taget="_blank" href="/search/BufferedImage/1.htm">BufferedImage</a>
<div>
package com.demo;
import java.awt.image.BufferedImage;
import java.io.File;
import java.io.IOException;
import javax.imageio.ImageIO;
/**
* @Description 读取一张图片改变大小
* @author FuJianyon</div>
</li>
<li><a href="/article/761.htm"
title="myeclipse集成svn（一针见血）" target="_blank">myeclipse集成svn（一针见血）</a>
<span class="text-muted">7454103</span>
<a class="tag" taget="_blank" href="/search/eclipse/1.htm">eclipse</a><a class="tag" taget="_blank" href="/search/SVN/1.htm">SVN</a><a class="tag" taget="_blank" href="/search/MyEclipse/1.htm">MyEclipse</a>
<div> &n</div>
</li>
<li><a href="/article/888.htm"
title="装箱与拆箱----autoboxing和unboxing" target="_blank">装箱与拆箱----autoboxing和unboxing</a>
<span class="text-muted">darkranger</span>
<a class="tag" taget="_blank" href="/search/J2SE/1.htm">J2SE</a>
<div>4.2 自动装箱和拆箱
基本数据(Primitive)类型的自动装箱(autoboxing)、拆箱(unboxing)是自J2SE 5.0开始提供的功能。虽然为您打包基本数据类型提供了方便，但提供方便的同时表示隐藏了细节，建议在能够区分基本数据类型与对象的差别时再使用。
4.2.1 autoboxing和unboxing
在Java中，所有要处理的东西几乎都是对象(Object)</div>
</li>
<li><a href="/article/1015.htm"
title="ajax传统的方式制作ajax" target="_blank">ajax传统的方式制作ajax</a>
<span class="text-muted">aijuans</span>
<a class="tag" taget="_blank" href="/search/Ajax/1.htm">Ajax</a>
<div>//这是前台的代码
<%@ page language="java" import="java.util.*" pageEncoding="UTF-8"%> <% String path = request.getContextPath(); String basePath = request.getScheme()+</div>
</li>
<li><a href="/article/1142.htm"
title="只用jre的eclipse是怎么编译java源文件的？" target="_blank">只用jre的eclipse是怎么编译java源文件的？</a>
<span class="text-muted">avords</span>
<a class="tag" taget="_blank" href="/search/java/1.htm">java</a><a class="tag" taget="_blank" href="/search/eclipse/1.htm">eclipse</a><a class="tag" taget="_blank" href="/search/jdk/1.htm">jdk</a><a class="tag" taget="_blank" href="/search/tomcat/1.htm">tomcat</a>
<div>eclipse只需要jre就可以运行开发java程序了，也能自动 编译java源代码，但是jre不是java的运行环境么，难道jre中也带有编译工具？ 还是eclipse自己实现的？谁能给解释一下呢问题补充：假设系统中没有安装jdk or jre，只在eclipse的目录中有一个jre，那么eclipse会采用该jre，问题是eclipse照样可以编译java源文件，为什么呢？
&nb</div>
</li>
<li><a href="/article/1269.htm"
title="前端模块化" target="_blank">前端模块化</a>
<span class="text-muted">bee1314</span>
<a class="tag" taget="_blank" href="/search/%E6%A8%A1%E5%9D%97%E5%8C%96/1.htm">模块化</a>
<div>背景： 前端JavaScript模块化，其实已经不是什么新鲜事了。但是很多的项目还没有真正的使用起来，还处于刀耕火种的野蛮生长阶段。 JavaScript一直缺乏有效的包管理机制，造成了大量的全局变量，大量的方法冲突。我们多么渴望有天能像Java（import），Python (import)，Ruby(require)那样写代码。在没有包管理机制的年代，我们是怎么避免所</div>
</li>
<li><a href="/article/1396.htm"
title="处理百万级以上的数据处理" target="_blank">处理百万级以上的数据处理</a>
<span class="text-muted">bijian1013</span>
<a class="tag" taget="_blank" href="/search/oracle/1.htm">oracle</a><a class="tag" taget="_blank" href="/search/sql/1.htm">sql</a><a class="tag" taget="_blank" href="/search/%E6%95%B0%E6%8D%AE%E5%BA%93/1.htm">数据库</a><a class="tag" taget="_blank" href="/search/%E5%A4%A7%E6%95%B0%E6%8D%AE/1.htm">大数据</a><a class="tag" taget="_blank" href="/search/%E6%9F%A5%E8%AF%A2/1.htm">查询</a>
<div>一.处理百万级以上的数据提高查询速度的方法： 1.应尽量避免在 where 子句中使用!=或<>操作符，否则将引擎放弃使用索引而进行全表扫描。
2.对查询进行优化，应尽量避免全表扫描，首先应考虑在 where 及 o</div>
</li>
<li><a href="/article/1523.htm"
title="mac 卸载 java 1.7 或更高版本" target="_blank">mac 卸载 java 1.7 或更高版本</a>
<span class="text-muted">征客丶</span>
<a class="tag" taget="_blank" href="/search/java/1.htm">java</a><a class="tag" taget="_blank" href="/search/OS/1.htm">OS</a>
<div>卸载 java 1.7 或更高
sudo rm -rf /Library/Internet\ Plug-Ins/JavaAppletPlugin.plugin
成功执行此命令后，还可以执行 java 与 javac 命令
sudo rm -rf /Library/PreferencePanes/JavaControlPanel.prefPane
成功执行此命令后，还可以执行 java </div>
</li>
<li><a href="/article/1650.htm"
title="【Spark六十一】Spark Streaming结合Flume、Kafka进行日志分析" target="_blank">【Spark六十一】Spark Streaming结合Flume、Kafka进行日志分析</a>
<span class="text-muted">bit1129</span>
<a class="tag" taget="_blank" href="/search/Stream/1.htm">Stream</a>
<div>第一步，Flume和Kakfa对接，Flume抓取日志，写到Kafka中
第二部，Spark Streaming读取Kafka中的数据，进行实时分析
本文首先使用Kakfa自带的消息处理（脚本）来获取消息，走通Flume和Kafka的对接 1. Flume配置
1. 下载Flume和Kafka集成的插件，下载地址：https://github.com/beyondj2ee/f</div>
</li>
<li><a href="/article/1777.htm"
title="Erlang vs TNSDL" target="_blank">Erlang vs TNSDL</a>
<span class="text-muted">bookjovi</span>
<a class="tag" taget="_blank" href="/search/erlang/1.htm">erlang</a>
<div>
TNSDL是Nokia内部用于开发电信交换软件的私有语言，是在SDL语言的基础上加以修改而成，TNSDL需翻译成C语言得以编译执行，TNSDL语言中实现了异步并行的特点，当然要完整实现异步并行还需要运行时动态库的支持，异步并行类似于Erlang的process（轻量级进程），TNSDL中则称之为hand，Erlang是基于vm(beam)开发，</div>
</li>
<li><a href="/article/1904.htm"
title="非常希望有一个预防疲劳的java软件, 预防过劳死和眼睛疲劳,大家一起努力搞一个" target="_blank">非常希望有一个预防疲劳的java软件, 预防过劳死和眼睛疲劳,大家一起努力搞一个</a>
<span class="text-muted">ljy325</span>
<a class="tag" taget="_blank" href="/search/%E4%BC%81%E4%B8%9A%E5%BA%94%E7%94%A8/1.htm">企业应用</a>
<div> 非常希望有一个预防疲劳的java软件，我看新闻和网站，国防科技大学的科学家累死了，太疲劳，老是加班，不休息，经常吃药，吃药根本就没用，根本原因是疲劳过度。我以前做java,那会公司垃圾，老想赶快学习到东西跳槽离开，搞得超负荷，不明理。深圳做软件开发经常累死人，总有不明理的人，有个软件提醒限制很好，可以挽救很多人的生命。
相关新闻：
（1）IT行业成五大疾病重灾区：过劳死平均37.9岁
</div>
</li>
<li><a href="/article/2031.htm"
title="读《研磨设计模式》-代码笔记-原型模式" target="_blank">读《研磨设计模式》-代码笔记-原型模式</a>
<span class="text-muted">bylijinnan</span>
<a class="tag" taget="_blank" href="/search/java/1.htm">java</a><a class="tag" taget="_blank" href="/search/%E8%AE%BE%E8%AE%A1%E6%A8%A1%E5%BC%8F/1.htm">设计模式</a>
<div>声明： 本文只为方便我个人查阅和理解，详细的分析以及源代码请移步 原作者的博客http://chjavach.iteye.com/
/**
* Effective Java 建议使用copy constructor or copy factory来代替clone()方法：
* 1.public Product copy(Product p){}
* 2.publi</div>
</li>
<li><a href="/article/2158.htm"
title="配置管理---svn工具之权限配置" target="_blank">配置管理---svn工具之权限配置</a>
<span class="text-muted">chenyu19891124</span>
<a class="tag" taget="_blank" href="/search/SVN/1.htm">SVN</a>
<div>今天花了大半天的功夫，终于弄懂svn权限配置。下面是今天收获的战绩。
安装完svn后就是在svn中建立版本库，比如我本地的是版本库路径是C:\Repositories\pepos。pepos是我的版本库。在pepos的目录结构
pepos
component
webapps
在conf里面的auth里赋予的权限配置为
[groups]</div>
</li>
<li><a href="/article/2285.htm"
title="浅谈程序员的数学修养" target="_blank">浅谈程序员的数学修养</a>
<span class="text-muted">comsci</span>
<a class="tag" taget="_blank" href="/search/%E8%AE%BE%E8%AE%A1%E6%A8%A1%E5%BC%8F/1.htm">设计模式</a><a class="tag" taget="_blank" href="/search/%E7%BC%96%E7%A8%8B/1.htm">编程</a><a class="tag" taget="_blank" href="/search/%E7%AE%97%E6%B3%95/1.htm">算法</a><a class="tag" taget="_blank" href="/search/%E9%9D%A2%E8%AF%95/1.htm">面试</a><a class="tag" taget="_blank" href="/search/%E6%8B%9B%E8%81%98/1.htm">招聘</a>
<div>
浅谈程序员的数学修养 </div>
</li>
<li><a href="/article/2412.htm"
title="批量执行 bulk collect与forall用法" target="_blank">批量执行 bulk collect与forall用法</a>
<span class="text-muted">daizj</span>
<a class="tag" taget="_blank" href="/search/oracle/1.htm">oracle</a><a class="tag" taget="_blank" href="/search/sql/1.htm">sql</a><a class="tag" taget="_blank" href="/search/bulk+collect/1.htm">bulk collect</a><a class="tag" taget="_blank" href="/search/forall/1.htm">forall</a>
<div>BULK COLLECT 子句会批量检索结果，即一次性将结果集绑定到一个集合变量中，并从SQL引擎发送到PL/SQL引擎。通常可以在SELECT INTO、
FETCH INTO以及RETURNING INTO子句中使用BULK COLLECT。本文将逐一描述BULK COLLECT在这几种情形下的用法。
有关FORALL语句的用法请参考：批量SQL之 F</div>
</li>
<li><a href="/article/2539.htm"
title="Linux下使用rsync最快速删除海量文件的方法" target="_blank">Linux下使用rsync最快速删除海量文件的方法</a>
<span class="text-muted">dongwei_6688</span>
<a class="tag" taget="_blank" href="/search/OS/1.htm">OS</a>
<div>1、先安装rsync：yum install rsync
2、建立一个空的文件夹：mkdir /tmp/test
3、用rsync删除目标目录：rsync --delete-before -a -H -v --progress --stats /tmp/test/ log/这样我们要删除的log目录就会被清空了，删除的速度会非常快。rsync实际上用的是替换原理，处理数十万个文件也是秒删。
</div>
</li>
<li><a href="/article/2666.htm"
title="Yii CModel中rules验证规格" target="_blank">Yii CModel中rules验证规格</a>
<span class="text-muted">dcj3sjt126com</span>
<a class="tag" taget="_blank" href="/search/rules/1.htm">rules</a><a class="tag" taget="_blank" href="/search/yii/1.htm">yii</a><a class="tag" taget="_blank" href="/search/validate/1.htm">validate</a>
<div>Yii cValidator主要用法分析：
yii验证rulesit 分类： Yii yii的rules验证 cValidator主要属性 attributes ,builtInValidators,enableClientValidation,message,on,safe,skipOnError
</div>
</li>
<li><a href="/article/2793.htm"
title="基于vagrant的redis主从实验" target="_blank">基于vagrant的redis主从实验</a>
<span class="text-muted">dcj3sjt126com</span>
<a class="tag" taget="_blank" href="/search/vagrant/1.htm">vagrant</a>
<div>平台: Mac
工具: Vagrant
系统: Centos6.5
实验目的: Redis主从
实现思路
制作一个基于sentos6.5, 已经安装好reids的box, 添加一个脚本配置从机, 然后作为后面主机从机的基础box
制作sentos6.5+redis的box
mkdir vagrant_redis
cd vagrant_</div>
</li>
<li><a href="/article/2920.htm"
title="Memcached(二)、Centos安装Memcached服务器" target="_blank">Memcached(二)、Centos安装Memcached服务器</a>
<span class="text-muted">frank1234</span>
<a class="tag" taget="_blank" href="/search/centos/1.htm">centos</a><a class="tag" taget="_blank" href="/search/memcached/1.htm">memcached</a>
<div>一、安装gcc
rpm和yum安装memcached服务器连接没有找到，所以我使用的是make的方式安装，由于make依赖于gcc，所以要先安装gcc
开始安装，命令如下，[color=red][b]顺序一定不能出错[/b][/color]：
建议可以先切换到root用户，不然可能会遇到权限问题：su root 输入密码......
rpm -ivh kernel-head</div>
</li>
<li><a href="/article/3047.htm"
title="Remove Duplicates from Sorted List" target="_blank">Remove Duplicates from Sorted List</a>
<span class="text-muted">hcx2013</span>
<a class="tag" taget="_blank" href="/search/remove/1.htm">remove</a>
<div>Given a sorted linked list, delete all duplicates such that each element appear only once.
For example,Given 1->1->2, return 1->2.Given 1->1->2->3->3, return&</div>
</li>
<li><a href="/article/3174.htm"
title="Spring4新特性——JSR310日期时间API的支持" target="_blank">Spring4新特性——JSR310日期时间API的支持</a>
<span class="text-muted">jinnianshilongnian</span>
<a class="tag" taget="_blank" href="/search/spring4/1.htm">spring4</a>
<div>Spring4新特性——泛型限定式依赖注入
Spring4新特性——核心容器的其他改进
Spring4新特性——Web开发的增强
Spring4新特性——集成Bean Validation 1.1(JSR-349)到SpringMVC
Spring4新特性——Groovy Bean定义DSL
Spring4新特性——更好的Java泛型操作API
Spring4新</div>
</li>
<li><a href="/article/3301.htm"
title="浅谈enum与单例设计模式" target="_blank">浅谈enum与单例设计模式</a>
<span class="text-muted">247687009</span>
<a class="tag" taget="_blank" href="/search/java/1.htm">java</a><a class="tag" taget="_blank" href="/search/%E5%8D%95%E4%BE%8B/1.htm">单例</a>
<div>在JDK1.5之前的单例实现方式有两种(懒汉式和饿汉式并无设计上的区别故看做一种)，两者同是私有构
造器，导出静态成员变量，以便调用者访问。
第一种
package singleton;
public class Singleton {
//导出全局成员
public final static Singleton INSTANCE = new S</div>
</li>
<li><a href="/article/3428.htm"
title="使用switch条件语句需要注意的几点" target="_blank">使用switch条件语句需要注意的几点</a>
<span class="text-muted">openwrt</span>
<a class="tag" taget="_blank" href="/search/c/1.htm">c</a><a class="tag" taget="_blank" href="/search/break/1.htm">break</a><a class="tag" taget="_blank" href="/search/switch/1.htm">switch</a>
<div>1. 当满足条件的case中没有break，程序将依次执行其后的每种条件（包括default）直到遇到break跳出
int main()
{
int n = 1;
switch(n) {
case 1:
printf("--1--\n");
default:
printf("defa</div>
</li>
<li><a href="/article/3555.htm"
title="配置Spring Mybatis JUnit测试环境的应用上下文" target="_blank">配置Spring Mybatis JUnit测试环境的应用上下文</a>
<span class="text-muted">schnell18</span>
<a class="tag" taget="_blank" href="/search/spring/1.htm">spring</a><a class="tag" taget="_blank" href="/search/mybatis/1.htm">mybatis</a><a class="tag" taget="_blank" href="/search/JUnit/1.htm">JUnit</a>
<div>Spring-test模块中的应用上下文和web及spring boot的有很大差异。主要试下来差异有：
单元测试的app context不支持从外部properties文件注入属性
@Value注解不能解析带通配符的路径字符串
解决第一个问题可以配置一个PropertyPlaceholderConfigurer的bean。
第二个问题的具体实例是：
</div>
</li>
<li><a href="/article/3682.htm"
title="Java 定时任务总结一" target="_blank">Java 定时任务总结一</a>
<span class="text-muted">tuoni</span>
<a class="tag" taget="_blank" href="/search/java/1.htm">java</a><a class="tag" taget="_blank" href="/search/spring/1.htm">spring</a><a class="tag" taget="_blank" href="/search/timer/1.htm">timer</a><a class="tag" taget="_blank" href="/search/quartz/1.htm">quartz</a><a class="tag" taget="_blank" href="/search/timertask/1.htm">timertask</a>
<div> Java定时任务总结 一.从技术上分类大概分为以下三种方式： 1.Java自带的java.util.Timer类，这个类允许你调度一个java.util.TimerTask任务; 说明： java.util.Timer定时器，实际上是个线程，定时执行TimerTask类 &</div>
</li>
<li><a href="/article/3809.htm"
title="一种防止用户生成内容站点出现商业广告以及非法有害等垃圾信息的方法" target="_blank">一种防止用户生成内容站点出现商业广告以及非法有害等垃圾信息的方法</a>
<span class="text-muted">yangshangchuan</span>
<a class="tag" taget="_blank" href="/search/rank/1.htm">rank</a><a class="tag" taget="_blank" href="/search/%E7%9B%B8%E4%BC%BC%E5%BA%A6%E8%AE%A1%E7%AE%97/1.htm">相似度计算</a><a class="tag" taget="_blank" href="/search/%E6%96%87%E6%9C%AC%E7%9B%B8%E4%BC%BC%E5%BA%A6/1.htm">文本相似度</a><a class="tag" taget="_blank" href="/search/%E8%AF%8D%E8%A2%8B%E6%A8%A1%E5%9E%8B/1.htm">词袋模型</a><a class="tag" taget="_blank" href="/search/%E4%BD%99%E5%BC%A6%E7%9B%B8%E4%BC%BC%E5%BA%A6/1.htm">余弦相似度</a>
<div>本文描述了一种在ITEYE博客频道上面出现的新型的商业广告形式及其应对方法，对于其他的用户生成内容站点类型也具有同样的适用性。
最近在ITEYE博客频道上面出现了一种新型的商业广告形式，方法如下：
1、注册多个账号（一般10个以上）。
2、从多个账号中选择一个账号，发表1-2篇博文</div>
</li>
</ul>
</div>
</div>
</div>
<div>
<div class="container">
<div class="indexes">
<strong>按字母分类：</strong>
<a href="/tags/A/1.htm" target="_blank">A</a><a href="/tags/B/1.htm" target="_blank">B</a><a href="/tags/C/1.htm" target="_blank">C</a><a
href="/tags/D/1.htm" target="_blank">D</a><a href="/tags/E/1.htm" target="_blank">E</a><a href="/tags/F/1.htm" target="_blank">F</a><a
href="/tags/G/1.htm" target="_blank">G</a><a href="/tags/H/1.htm" target="_blank">H</a><a href="/tags/I/1.htm" target="_blank">I</a><a
href="/tags/J/1.htm" target="_blank">J</a><a href="/tags/K/1.htm" target="_blank">K</a><a href="/tags/L/1.htm" target="_blank">L</a><a
href="/tags/M/1.htm" target="_blank">M</a><a href="/tags/N/1.htm" target="_blank">N</a><a href="/tags/O/1.htm" target="_blank">O</a><a
href="/tags/P/1.htm" target="_blank">P</a><a href="/tags/Q/1.htm" target="_blank">Q</a><a href="/tags/R/1.htm" target="_blank">R</a><a
href="/tags/S/1.htm" target="_blank">S</a><a href="/tags/T/1.htm" target="_blank">T</a><a href="/tags/U/1.htm" target="_blank">U</a><a
href="/tags/V/1.htm" target="_blank">V</a><a href="/tags/W/1.htm" target="_blank">W</a><a href="/tags/X/1.htm" target="_blank">X</a><a
href="/tags/Y/1.htm" target="_blank">Y</a><a href="/tags/Z/1.htm" target="_blank">Z</a><a href="/tags/0/1.htm" target="_blank">其他</a>
</div>
</div>
</div>
<footer id="footer" class="mb30 mt30">
<div class="container">
<div class="footBglm">
<a target="_blank" href="/">首页</a> -
<a target="_blank" href="/custom/about.htm">关于我们</a> -
<a target="_blank" href="/search/Java/1.htm">站内搜索</a> -
<a target="_blank" href="/sitemap.txt">Sitemap</a> -
<a target="_blank" href="/custom/delete.htm">侵权投诉</a>
</div>
<div class="copyright">版权所有 IT知识库 CopyRight © 2000-2050 E-COM-NET.COM , All Rights Reserved.
<!-- <a href="https://beian.miit.gov.cn/" rel="nofollow" target="_blank">京ICP备09083238号</a><br>-->
</div>
</div>
</footer>
<!-- 代码高亮 -->
<script type="text/javascript" src="/static/syntaxhighlighter/scripts/shCore.js"></script>
<script type="text/javascript" src="/static/syntaxhighlighter/scripts/shLegacy.js"></script>
<script type="text/javascript" src="/static/syntaxhighlighter/scripts/shAutoloader.js"></script>
<link type="text/css" rel="stylesheet" href="/static/syntaxhighlighter/styles/shCoreDefault.css"/>
<script type="text/javascript" src="/static/syntaxhighlighter/src/my_start_1.js"></script>
</body>
</html>