自从ChatGPT发布以来,国内外的开源大模型如雨后春笋般成长,但是对于很多企业和个人从头训练预训练模型不太现实,即使微调开源大模型也捉襟见肘,那么直接部署这些开源大模型服务于企业业务将会有很大的前景,本文将介绍七中主流的LLM推理和服务开源库。
下面首先来总结一下这些框架的特点,如下表所示:
LLM推理有很多框架,各有其特点,下面分别介绍一下表中七个框架的关键点:
vLLM[1]:适用于大批量Prompt输入,并对推理速度要求高的场景;
Text generation inference[2]:依赖HuggingFace模型,并且不需要为核心模型增加多个adapter的场景;
CTranslate2[3]:可在CPU上进行推理;
OpenLLM[4]:为核心模型添加adapter并使用HuggingFace Agents,尤其是不完全依赖PyTorch;
Ray Serve[5]:稳定的Pipeline和灵活的部署,它最适合更成熟的项目;
MLC LLM[6]:可在客户端(边缘计算)(例如,在Android或iPhone平台上)本地部署LLM;
DeepSpeed-MII[7]:使用DeepSpeed库来部署LLM;
下面我们在内存容量为40GB的A100 GPU上,并且使用LLaMA-1 13b模型(因为列表中的所有库都支持它)进行七个部署框架的对比。
vLLM的吞吐量比HuggingFace Transformers(HF)高14x-24倍,比HuggingFace Text Generation Inference(TGI)高2.2x-2.5倍。
# pip install vllm
from vllm import LLM, SamplingParams
prompts = [
"Funniest joke ever:",
"The capital of France is",
"The future of AI is",
]
sampling_params = SamplingParams(temperature=0.95, top_p=0.95, max_tokens=200)
llm = LLM(model="huggyllama/llama-13b")
outputs = llm.generate(prompts, sampling_params)
for output in outputs:
prompt = output.prompt
generated_text = output.outputs[0].text
print(f"Prompt: {prompt!r}, Generated text: {generated_text!r}")
# Start the server:
python -m vllm.entrypoints.api_server --env MODEL_NAME=huggyllama/llama-13b
# Query the model in shell:
curl http://localhost:8000/generate \
-d '{
"prompt": "Funniest joke ever:",
"n": 1,
"temperature": 0.95,
"max_tokens": 200
}'
这是LLM推理最快的库。得益于其内部优化,它显著优于竞争对手。尽管如此,它在支持有限范围的模型方面确实存在弱点。
使用vLLM的开发路线可以参考:https://github.com/vllm-project/vllm/issues/244
Text generation inference是用于文本生成推断的Rust、Python和gRPC服务器,在HuggingFace中已有LLM 推理API使用。
mkdir data
docker run --gpus all --shm-size 1g -p 8080:80 \
-v data:/data ghcr.io/huggingface/text-generation-inference:0.9 \
--model-id huggyllama/llama-13b \
--num-shard 1
# pip install text-generation
from text_generation import Client
client = Client("http://127.0.0.1:8080")
prompt = "Funniest joke ever:"
print(client.generate(prompt, max_new_tokens=17 temperature=0.95).generated_text)
使用Text generation inference的开发路线可以参考:https://github.com/huggingface/text-generation-inference/issues/232
CTranslate2是一个C++和Python库,用于使用Transformer模型进行高效推理。
pip install -qqq transformers ctranslate2
# The model should be first converted into the CTranslate2 model format:
ct2-transformers-converter --model huggyllama/llama-13b --output_dir llama-13b-ct2 --force
import ctranslate2
import transformers
generator = ctranslate2.Generator("llama-13b-ct2", device="cuda", compute_type="float16")
tokenizer = transformers.AutoTokenizer.from_pretrained("huggyllama/llama-13b")
prompt = "Funniest joke ever:"
tokens = tokenizer.convert_ids_to_tokens(tokenizer.encode(prompt))
results = generator.generate_batch(
[tokens],
sampling_topk=1,
max_length=200,
)
tokens = results[0].sequences_ids[0]
output = tokenizer.decode(tokens)
print(output)
在DeepSpeed支持下,DeepSpeed-MII可以进行低延迟和高通量推理。
# DON'T INSTALL USING pip install deepspeed-mii
# git clone https://github.com/microsoft/DeepSpeed-MII.git
# git reset --hard 60a85dc3da5bac3bcefa8824175f8646a0f12203
# cd DeepSpeed-MII && pip install .
# pip3 install -U deepspeed
# ... and make sure that you have same CUDA versions:
# python -c "import torch;print(torch.version.cuda)" == nvcc --version
import mii
mii_configs = {
"dtype": "fp16",
'max_tokens': 200,
'tensor_parallel': 1,
"enable_load_balancing": False
}
mii.deploy(task="text-generation",
model="huggyllama/llama-13b",
deployment_name="llama_13b_deployment",
mii_config=mii_configs)
import mii
generator = mii.mii_query_handle("llama_13b_deployment")
result = generator.query(
{"query": ["Funniest joke ever:"]},
do_sample=True,
max_new_tokens=200
)
print(result)
OpenLLM是一个用于在生产中操作大型语言模型(LLM)的开放平台。
pip install openllm scipy
openllm start llama --model-id huggyllama/llama-13b \
--max-new-tokens 200 \
--temperature 0.95 \
--api-workers 1 \
--workers-per-resource 1
import openllm
client = openllm.client.HTTPClient('http://localhost:3000')
print(client.query("Funniest joke ever:"))
Ray Serve是一个可扩展的模型服务库,用于构建在线推理API。Serve与框架无关,因此可以使用一个工具包来为深度学习模型的所有内容提供服务。
# pip install ray[serve] accelerate>=0.16.0 transformers>=4.26.0 torch starlette pandas
# ray_serve.py
import pandas as pd
import ray
from ray import serve
from starlette.requests import Request
@serve.deployment(ray_actor_options={"num_gpus": 1})
class PredictDeployment:
def __init__(self, model_id: str):
from transformers import AutoModelForCausalLM, AutoTokenizer
import torch
self.model = AutoModelForCausalLM.from_pretrained(
model_id,
torch_dtype=torch.float16,
device_map="auto",
)
self.tokenizer = AutoTokenizer.from_pretrained(model_id)
def generate(self, text: str) -> pd.DataFrame:
input_ids = self.tokenizer(text, return_tensors="pt").input_ids.to(
self.model.device
)
gen_tokens = self.model.generate(
input_ids,
temperature=0.9,
max_length=200,
)
return pd.DataFrame(
self.tokenizer.batch_decode(gen_tokens), columns=["responses"]
)
async def __call__(self, http_request: Request) -> str:
json_request: str = await http_request.json()
return self.generate(prompt["text"])
deployment = PredictDeployment.bind(model_id="huggyllama/llama-13b")
# then run from CLI command:
# serve run ray_serve:deployment
import requests
sample_input = {"text": "Funniest joke ever:"}
output = requests.post("http://localhost:8000/", json=[sample_input]).json()
print(output)
如果需要最适合生产的解决方案,而不仅仅是深度学习,Ray Serve是一个不错的选择。它最适合于可用性、可扩展性和可观察性非常重要的企业。此外,还可以使用其庞大的生态系统进行数据处理、模型训练、微调和服务。最后,从OpenAI到Shopify和Instacart等公司都在使用它。
LLM的机器学习编译(MLC LLM)是一种通用的部署解决方案,它使LLM能够利用本机硬件加速在消费者设备上高效运行。
# 1. Make sure that you have python >= 3.9
# 2. You have to run it using conda:
conda create -n mlc-chat-venv -c mlc-ai -c conda-forge mlc-chat-nightly
conda activate mlc-chat-venv
# 3. Then install package:
pip install --pre --force-reinstall mlc-ai-nightly-cu118 \
mlc-chat-nightly-cu118 \
-f https://mlc.ai/wheels
# 4. Download the model weights from HuggingFace and binary libraries:
git lfs install && mkdir -p dist/prebuilt && \
git clone https://github.com/mlc-ai/binary-mlc-llm-libs.git dist/prebuilt/lib && \
cd dist/prebuilt && \
git clone https://huggingface.co/huggyllama/llama-13b dist/ && \
cd ../..
# 5. Run server:
python -m mlc_chat.rest --device-name cuda --artifact-path dist
import requests
payload = {
"model": "lama-30b",
"messages": [{"role": "user", "content": "Funniest joke ever:"}],
"stream": False
}
r = requests.post("http://127.0.0.1:8000/v1/chat/completions", json=payload)
print(r.json()['choices'][0]['message']['content'])
如果需要在iOS或Android设备上部署应用程序,这个库正是你所需要的。它将允许您快速地以本机方式编译模型并将其部署到设备上。但是,如果需要一个高负载的服务器,不建议选择这个框架。
[1] https://github.com/vllm-project/vllm
[2] https://github.com/huggingface/text-generation-inference
[3] https://github.com/OpenNMT/CTranslate2
[4] https://github.com/bentoml/OpenLLM
[5] https://docs.ray.io/en/latest/serve/index.html
[6] https://github.com/mlc-ai/mlc-llm
[7] https://github.com/microsoft/DeepSpeed-MII
[8] https://github.com/microsoft/DeepSpeed
[9] https://www.anyscale.com/blog/continuous-batching-llm-inference
[10] https://vllm.ai/
[11] https://huggingface.co/docs/transformers/main_classes/agent
[12] https://github.com/TimDettmers/bitsandbytes
[13] https://arxiv.org/abs/2210.17323
[14] https://github.com/bentoml/Yatai
[15] https://arxiv.org/abs/2212.09720