Mask R-CNN训练自己的数据集在win10上的踩坑全过程：CUDA9.0+CUDNN7.1.4+Tensorflow-gpu1.9.0+keras-gpu2.2.4

基础配置

1. 首先你需要在win10上下载Git（用于我们在github上面下载源码）和MinGW（方便我们在win10上也能用linux的make操作命令）。
2. 接着你要下载cuda9.0和cudnn7.1来绑定你的windows的Nvidia
3. 接着你需要在win10上面安装anaconda3（切记，python用的是3.6+，目前的tesorflow-gpu只能匹配这个）
4. 然后在现有的base环境下(或者配置新环境），按照顺序依次用conda install cudatoolkit==9.0->conda install cudnn==7.1.4->conda install tensorflow-gpu==1.9.0->conda install keras-gpu==2.2.4
5. 然后记住要装一些其他需要的包，比如（matplotlib、openCV、cython、scikit-image、imgaug等等），总之就是按照提示conda install或者pip install

前期需要下载的内容

1. 首先大家用在github上把mask r-cnn的项目下载到本地；
2. 然后要把cocoapi下载到本地；
3. 接着是把预权重mask_rcnn_coco.h5放到mask r-cnn文件下面的新建文件夹logs下面；
4. 最后是打开anaconda prompt，cd到cocoapi/PythonAPI的目录，一次输入make和make -j8；
5. 把PythonAPI下面更新的pycocotools放到mask r-cnn文件的samples/coco/里；

需要注意的地方 

1. mask r-cnn在训练的时候，gpu显存会被一直占用，但是利用率只会偶尔上升是正常现象，大部分时间gpu利用率还是比较低，只要gpu显存被占用就说明后方是在用gpu加速的。
2. 同时cpu内存也会被占用一些，利用率大概在百分之三十多。

先附上mask-rcnn的官方测试代码：

import os
import sys
import random
import math
import numpy as np
import skimage.io
import matplotlib
import matplotlib.pyplot as plt
import cv2
import time

ROOT_DIR = os.path.abspath("G:\Mask_RCNN")

sys.path.append(ROOT_DIR)
from mrcnn import utils
import mrcnn.model as modellib
from mrcnn import visualize

sys.path.append(os.path.join(ROOT_DIR, "samples/coco/"))
import coco

MODEL_DIR = os.path.join(ROOT_DIR, "logs")

COCO_MODEL_PATH = os.path.join(MODEL_DIR ,"mask_rcnn_coco.h5")

if not os.path.exists(COCO_MODEL_PATH):
    utils.download_trained_weights(COCO_MODEL_PATH)
    print("cuiwei***********************")

IMAGE_DIR = os.path.join(ROOT_DIR, "images")

class InferenceConfig(coco.CocoConfig):
    GPU_COUNT = 1
    IMAGES_PER_GPU = 1
    
config = InferenceConfig()
config.display()

model = modellib.MaskRCNN(mode="inference", model_dir=MODEL_DIR, config=config)

#model.load_weights(COCO_MODEL_PATH, by_name=True)
model.load_weights(COCO_MODEL_PATH, by_name=True, exclude=[ "mrcnn_class_logits", "mrcnn_bbox_fc"])
# =============================================================================
# class_names = ['BG', 'person', 'bicycle', 'car', 'motorcycle', 'airplane',
#                'bus', 'train', 'truck', 'boat', 'traffic light',
#                'fire hydrant', 'stop sign', 'parking meter', 'bench', 'bird',
#                'cat', 'dog', 'horse', 'sheep', 'cow', 'elephant', 'bear',
#                'zebra', 'giraffe', 'backpack', 'umbrella', 'handbag', 'tie',
#                'suitcase', 'frisbee', 'skis', 'snowboard', 'sports ball',
#                'kite', 'baseball bat', 'baseball glove', 'skateboard',
#                'surfboard', 'tennis racket', 'bottle', 'wine glass', 'cup',
#                'fork', 'knife', 'spoon', 'bowl', 'banana', 'apple',
#                'sandwich', 'orange', 'broccoli', 'carrot', 'hot dog', 'pizza',
#                'donut', 'cake', 'chair', 'couch', 'potted plant', 'bed',
#                'dining table', 'toilet', 'tv', 'laptop', 'mouse', 'remote',
#                'keyboard', 'cell phone', 'microwave', 'oven', 'toaster',
#                'sink', 'refrigerator', 'book', 'clock', 'vase', 'scissors',
#                'teddy bear', 'hair drier', 'toothbrush']
# =============================================================================
class_names = ['BG', 'person']
#file_names=next(os.walk(IMAGE_DIR))[2]
#image=skimage.io.imread(os.path.join(IMAGE_DIR,random.choice(file_names)))
image= cv2.imread("G:\\Mask_RCNN\\images\\25691390_f9944f61b5_z.jpg")

#cap = cv2.VideoCapture(0)
i=0
while(i<1):
    i+=1
#    ret, frame = cap.read()
    
    start =time.clock()
#    results = model.detect([frame], verbose=1)
    results=model.detect([image],verbose=1)
    
    r = results[0]
    
#    visualize.display_instances(frame, r['rois'], r['masks'], r['class_ids'], class_names, r['scores'])
    visualize.display_instances(image,r['rois'],r['masks'],r['class_ids'],class_names,r['scores'])
    end = time.clock()
    print(end-start)
    if cv2.waitKey(1) & 0xFF == ord('q'):
        break
    
#cap.release()
cv2.destroyAllWindows()

在测试代码中会踩到的坑

Traceback (most recent call last):

  File "", line 1, in 
    runfile('G:/labelme/test.py', wdir='G:/labelme')

  File "C:\Users\34905\Anaconda3\envs\cv2\lib\site-packages\spyder_kernels\customize\spydercustomize.py", line 704, in runfile
    execfile(filename, namespace)

  File "C:\Users\34905\Anaconda3\envs\cv2\lib\site-packages\spyder_kernels\customize\spydercustomize.py", line 108, in execfile
    exec(compile(f.read(), filename, 'exec'), namespace)

  File "G:/labelme/test.py", line 48, in 
    model.load_weights(COCO_MODEL_PATH, by_name=True)

  File "G:\Mask_RCNN\mrcnn\model.py", line 2131, in load_weights
    saving.load_weights_from_hdf5_group_by_name(f, layers)

  File "C:\Users\34905\Anaconda3\envs\cv2\lib\site-packages\keras\engine\saving.py", line 1149, in load_weights_from_hdf5_group_by_name
    str(weight_values[i].shape) + '.')

ValueError: Layer #389 (named "mrcnn_bbox_fc"), weight  has shape (1024, 8), but the saved weight has shape (1024, 324).

如果你不想测试coco里默认的81类，只想测试2类，那一定记住要把model.load_weights(COCO_MODEL_PATH, by_name=True)改为model.load_weights(COCO_MODEL_PATH, by_name=True, exclude=["mrcnn_class_logits", "mrcnn_bbox_fc","mrcnn_bbox", "mrcnn_mask"])

附上训练代码

import os
import sys
import random
import math
import re
import time
import numpy as np
import cv2
import matplotlib
import matplotlib.pyplot as plt
import tensorflow as tf

#ROOT_DIR = os.getcwd()
ROOT_DIR = os.path.abspath("G:\Mask_RCNN")
sys.path.append(ROOT_DIR)

from mrcnn.config import Config
#import utils
from mrcnn import model as modellib,utils
from mrcnn import visualize
import yaml
from mrcnn.model import log
from PIL import Image


MODEL_DIR = os.path.join(ROOT_DIR, "logs")

iter_num=0

COCO_MODEL_PATH = os.path.join(MODEL_DIR, "mask_rcnn_coco.h5")

if not os.path.exists(COCO_MODEL_PATH):
    utils.download_trained_weights(COCO_MODEL_PATH)
    
class ShapesConfig(Config):
     NAME = "pods"
    
     GPU_COUNT = 1
     
     IMAGES_PER_GPU = 2
    
     NUM_CLASSES = 1 + 1
     
     IMAGE_MIN_DIM = 320
     IMAGE_MAX_DIM = 384
     
#     RPN_ANCHOR_SCALES = (8 * 6, 16 * 6, 32 * 6, 64 * 6, 128 * 6)
     RPN_ANCHOR_SCALES = (32, 64, 128, 256, 512)
     
     RPN_ANCHOR_RATIOS = [0.5, 1, 2]
     
     RPN_ANCHOR_STRIDE = 1
     RPN_NMS_THRESHOLD = 0.7
     RPN_TRAIN_ANCHORS_PER_IMAGE = 256
     
     TRAIN_ROIS_PER_IMAGE = 100
     
     STEPS_PER_EPOCH = 28
     
     MAX_GT_INSTANCES = 100
     
     VALIDATION_STEPS = 5
config = ShapesConfig()
config.display()

class DrugDataset(utils.Dataset):
    def get_obj_index(self, image):
        n = np.max(image)
        return n
    
    def from_yaml_get_class(self, image_id):
        info = self.image_info[image_id]
        with open(info['yaml_path']) as f:
            temp = yaml.load(f.read())
            labels = temp['label_names']
            del labels[0]
        return labels
    
    def draw_mask(self, num_obj, mask, image,image_id):
        info = self.image_info[image_id]
        for index in range(num_obj):
            for i in range(info['width']):
                for j in range(info['height']):
                    at_pixel = image.getpixel((i, j))
                    if at_pixel == index + 1:
                        mask[j, i, index] = 1
        return mask
    
    def load_shapes(self, count, img_folder, mask_folder, imglist, dataset_root_path):
        self.add_class("shapes", 1, "pod")
        for i in range(count):
            filestr = imglist[i].split(".")[0]
            mask_path = mask_folder + "/" + filestr + ".png"
            yaml_path = dataset_root_path + "labelme_json/" + filestr + "_json/info.yaml"
            print(dataset_root_path + "labelme_json/" + filestr + "_json/img.png")
            cv_img = cv2.imread(dataset_root_path + "labelme_json/" + filestr + "_json/img.png")
            
            self.add_image("shapes", image_id=i, path=img_folder + "/" + imglist[i],
                           width=cv_img.shape[1], height=cv_img.shape[0], mask_path=mask_path, yaml_path=yaml_path)

    def load_mask(self, image_id):
        global iter_num
        print("image_id",image_id)
        info = self.image_info[image_id]
        count = 1  # number of object
        img = Image.open(info['mask_path'])
        num_obj = self.get_obj_index(img)
        mask = np.zeros([info['height'], info['width'], num_obj], dtype=np.uint8)
        mask = self.draw_mask(num_obj, mask, img,image_id)
        occlusion = np.logical_not(mask[:, :, -1]).astype(np.uint8)
        for i in range(count - 2, -1, -1):
            mask[:, :, i] = mask[:, :, i] * occlusion
            
            occlusion = np.logical_and(occlusion, np.logical_not(mask[:, :, i]))
        labels = []
        labels = self.from_yaml_get_class(image_id)
        labels_form = []
        for i in range(len(labels)):
            if labels[i].find("pod") != -1:
                # print "box"
                labels_form.append("pod")
        class_ids = np.array([self.class_names.index(s) for s in labels_form])
        return mask, class_ids.astype(np.int32)

def get_ax(rows=1, cols=1, size=8):
    _, ax = plt.subplots(rows, cols, figsize=(size * cols, size * rows))
    return ax

#dataset_root_path="train_data/"
dataset_root_path = os.path.join(ROOT_DIR, "train_data/")
img_folder = dataset_root_path + "pic"
mask_folder = dataset_root_path + "cv2_mask"
imglist = os.listdir(img_folder)
count = len(imglist)

dataset_train = DrugDataset()
dataset_train.load_shapes(count, img_folder, mask_folder, imglist,dataset_root_path)
dataset_train.prepare()

dataset_val = DrugDataset()
dataset_val.load_shapes(56, img_folder, mask_folder, imglist,dataset_root_path)
dataset_val.prepare()

model = modellib.MaskRCNN(mode="training", config=config,model_dir=MODEL_DIR)

init_with = "coco"

if init_with == "imagenet":
    model.load_weights(model.get_imagenet_weights(), by_name=True)
elif init_with == "coco":
    model.load_weights(COCO_MODEL_PATH, by_name=True,exclude=["mrcnn_class_logits", "mrcnn_bbox_fc","mrcnn_bbox", "mrcnn_mask"])
elif init_with == "last":
    model.load_weights(model.find_last()[1], by_name=True)
    
#model.train(dataset_train, dataset_val,learning_rate=config.LEARNING_RATE,epochs=1,layers='heads')
model.train(dataset_train, dataset_val,learning_rate=config.LEARNING_RATE,epochs=20,layers='heads')

model.train(dataset_train, dataset_val,learning_rate=config.LEARNING_RATE / 10,epochs=20,layers="all")
#model.train(dataset_train, dataset_val,learning_rate=config.LEARNING_RATE / 10,epochs=1,layers="all")

训练代码中遇到的坑

 

Traceback (most recent call last):

  File "", line 1, in 
    runfile('D:/Mask R-CNN/train.py', wdir='D:/Mask R-CNN')

  File "D:\Anaconda3\Anaconda3-5.3.0\envs\cv2\lib\site-packages\spyder_kernels\customize\spydercustomize.py", line 704, in runfile
    execfile(filename, namespace)

  File "D:\Anaconda3\Anaconda3-5.3.0\envs\cv2\lib\site-packages\spyder_kernels\customize\spydercustomize.py", line 108, in execfile
    exec(compile(f.read(), filename, 'exec'), namespace)

  File "D:/Mask R-CNN/train.py", line 158, in 
    model.load_weights(COCO_MODEL_PATH, by_name=True,exclude=["mrcnn_class_logits", "mrcnn_bbox_fc","mrcnn_bbox", "mrcnn_mask"])

  File "D:\Mask R-CNN\mrcnn\model.py", line 2131, in load_weights
    saving.load_weights_from_hdf5_group_by_name(f, layers)

  File "D:\Anaconda3\Anaconda3-5.3.0\envs\cv2\lib\site-packages\keras\engine\saving.py", line 1104, in load_weights_from_hdf5_group_by_name
    g = f[name]

  File "h5py\_objects.pyx", line 54, in h5py._objects.with_phil.wrapper

  File "h5py\_objects.pyx", line 55, in h5py._objects.with_phil.wrapper

  File "D:\Anaconda3\Anaconda3-5.3.0\envs\cv2\lib\site-packages\h5py\_hl\group.py", line 177, in __getitem__
    oid = h5o.open(self.id, self._e(name), lapl=self._lapl)

  File "h5py\_objects.pyx", line 54, in h5py._objects.with_phil.wrapper

  File "h5py\_objects.pyx", line 55, in h5py._objects.with_phil.wrapper

  File "h5py\h5o.pyx", line 190, in h5py.h5o.open

KeyError: 'Unable to open object (wrong B-tree signature)'

 

 

 

这个bug是我目前遇到最难察觉问题的bug，而且网上关于这个bug的解决方法很少，经过我仔细的排查，我发现是我的mask_rcnn_coco.h5文件有问题。我重新下载后，就解决了这个bug，建议大家在重新搭建平台的时候，不要用上一个平台已经用过的预权重 mask_rcnn_coco.h5，会发生很多不可知的错误。最好还是重新去网上下载一个。

测试自己的训练代码

import os
import sys
import random
import math
import numpy as np
import skimage.io
import matplotlib
import matplotlib.pyplot as plt
import cv2
import time

ROOT_DIR = os.path.abspath("G:\Mask_RCNN")

sys.path.append(ROOT_DIR)

from mrcnn.config import Config
from datetime import datetime
from mrcnn import utils
import mrcnn.model as modellib
from mrcnn import visualize

sys.path.append(os.path.join(ROOT_DIR, "samples/coco/"))
from samples.coco import coco

MODEL_DIR = os.path.join(ROOT_DIR, "logs/shapes20181122T2257")
#MODEL_DIR = os.path.join(ROOT_DIR, "logs")

COCO_MODEL_PATH = os.path.join(MODEL_DIR, "mask_rcnn_shapes_0011.h5")
#COCO_MODEL_PATH = os.path.join(MODEL_DIR, "mask_rcnn_coco.h5")

IMAGE_DIR = os.path.join(ROOT_DIR, "train_data/pic")
#IMAGE_DIR = os.path.join(ROOT_DIR, "images")

class ShapesConfig(Config):
    NAME = "shapes"
    
    GPU_COUNT = 1
    IMAGES_PER_GPU = 1
    
    NUM_CLASSES = 1 + 1
    
    IMAGE_MIN_DIM = 320
    IMAGE_MAX_DIM = 384
    
    RPN_ANCHOR_SCALES = (8 * 6, 16 * 6, 32 * 6, 64 * 6, 128 * 6)
    
    TRAIN_ROIS_PER_IMAGE =100
    
    STEPS_PER_EPOCH = 10
    
    VALIDATION_STEPS = 10
    
class InferenceConfig(ShapesConfig):
    
    GPU_COUNT = 1
    IMAGES_PER_GPU = 1
    
config = InferenceConfig()

model = modellib.MaskRCNN(mode="inference", model_dir=MODEL_DIR, config=config)

model = modellib.MaskRCNN(mode="inference", model_dir=MODEL_DIR, config=config)

model.load_weights(COCO_MODEL_PATH, by_name=True)

class_names = ['BG', 'pod']
i=1
while(i<2):
    i=i+1
    file_names = next(os.walk(IMAGE_DIR))[2]
#    image = skimage.io.imread(os.path.join(IMAGE_DIR, random.choice(file_names)))
    image = skimage.io.imread(os.path.join(IMAGE_DIR, "1001-01b.jpg"))
    a=datetime.now()

    results = model.detect([image], verbose=1)
    b=datetime.now()

    print("time:",(b-a).seconds)
    r = results[0]
    visualize.display_instances(image, r['rois'], r['masks'], r['class_ids'], class_names, r['scores'])

同样的在Linux的ubutu16.04上测试我的代码的时候遇到了这个问题。

 

遇到这个问题大家不要慌，这是因为大家设置了多线程，但是线程不同步造成的。大家耐心等一会，就会出现loss了。同时如果大家不希望出现多线程，大家可以改为单线程。更改方法如下：在Mask RCNN\mrcnn\model.py中

 self.keras_model.fit_generator(
            train_generator,
            initial_epoch=self.epoch,
            epochs=epochs,
            steps_per_epoch=self.config.STEPS_PER_EPOCH,
            callbacks=callbacks,
            validation_data=val_generator,
            validation_steps=self.config.VALIDATION_STEPS,
            max_queue_size=100,
            workers=workers,
            use_multiprocessing=True,
#            use_multiprocessing=False,
        )

大家如果只是想要把多线程改为单线程，就要把use_multiprocessing=False,同时要让workers=1。因为单线程的情况下让workers大于1会报错！

有什么其他问题欢迎大家在下方留言私信我！