detectron2训练报错

训练时报错报错

-- Process 2 terminated with the following error:
Traceback (most recent call last):
  File "/home/hxw/anaconda3/envs/xinwen/lib/python3.6/site-packages/torch/multiprocessing/spawn.py", line 19, in _wrap
    fn(i, *args)
  File "/home/hxw/detectron2/detectron2/engine/launch.py", line 84, in _distributed_worker
    main_func(*args)
  File "/home/hxw/detectron2/tools/train_net.py", line 149, in main
    return trainer.train()
  File "/home/hxw/detectron2/detectron2/engine/defaults.py", line 330, in train
    super().train(self.start_iter, self.max_iter)
  File "/home/hxw/detectron2/detectron2/engine/train_loop.py", line 132, in train
    self.run_step()
  File "/home/hxw/detectron2/detectron2/engine/train_loop.py", line 206, in run_step
    data = next(self._data_loader_iter)
  File "/home/hxw/anaconda3/envs/xinwen/lib/python3.6/site-packages/torch/utils/data/dataloader.py", line 801, in __next__
    return self._process_data(data)
  File "/home/hxw/anaconda3/envs/xinwen/lib/python3.6/site-packages/torch/utils/data/dataloader.py", line 846, in _process_data
    data.reraise()
  File "/home/hxw/anaconda3/envs/xinwen/lib/python3.6/site-packages/torch/_utils.py", line 385, in reraise
    raise self.exc_type(msg)
OSError: Caught OSError in DataLoader worker process 3.
Original Traceback (most recent call last):
  File "/home/hxw/anaconda3/envs/xinwen/lib/python3.6/site-packages/torch/utils/data/_utils/worker.py", line 178, in _worker_loop
    data = fetcher.fetch(index)
  File "/home/hxw/anaconda3/envs/xinwen/lib/python3.6/site-packages/torch/utils/data/_utils/fetch.py", line 44, in fetch
    data = [self.dataset[idx] for idx in possibly_batched_index]
  File "/home/hxw/anaconda3/envs/xinwen/lib/python3.6/site-packages/torch/utils/data/_utils/fetch.py", line 44, in <listcomp>
    data = [self.dataset[idx] for idx in possibly_batched_index]
  File "/home/hxw/detectron2/detectron2/data/common.py", line 39, in __getitem__
    data = self._map_func(self._dataset[cur_idx])
  File "/home/hxw/detectron2/detectron2/data/dataset_mapper.py", line 72, in __call__
    image = utils.read_image(dataset_dict["file_name"], format=self.img_format)
  File "/home/hxw/detectron2/detectron2/data/detection_utils.py", line 53, in read_image
    image = image.convert(conversion_format)
  File "/home/hxw/anaconda3/envs/xinwen/lib/python3.6/site-packages/PIL/Image.py", line 930, in convert
    self.load()
  File "/home/hxw/anaconda3/envs/xinwen/lib/python3.6/site-packages/PIL/ImageFile.py", line 249, in load
    "(%d bytes not processed)" % len(b)
OSError: image file is truncated (7 bytes not processed)

单独读取图片时报错

“premature end of JPEG file”

原因

正常JPEG文件结束标识EOI，文件尾2个字节：0xff，0xd9
异常文件则不是这两个标识符，可以自行查看。

解决办法

代码需要的模块

import os
import cv2
from io import BytesIO
import skimage
import numpy as np
from PIL import Image

• 方式一

def is_broken_img(img_path):
    num = 0
    index = 0
    for file_ in os.listdir(img_path):
        img_ = os.path.join(img_path, file_)
        with open(img_, 'rb') as f:
            check_chars = f.read()[-2:]
        if check_chars != b'\xff\xd9':
            print("%s image is broken. "%file_)
            num += 1
        if index%2000 == 0:
            print("index of ok image: ", index)
        index += 1
    print("number of broken images: ",num)

应用举例：

is_broken_img("val2017")

• 方式二

def is_valid_file(img_path):
    num = 0
    index = 0
    try:
        for file_ in os.listdir(img_path):
            imgfile = None
            img_ = os.path.join(img_path, file_)
            with open(img_, 'rb') as f:
                buff = BytesIO()
                buff.write(f.read())
                buff.seek(0)
                temp_img = np.array(Image.open(buff), dtype=np.uint8)
                imgfile = cv2.cvtColor(temp_img, cv2.COLOR_RGB2BGR)
                buff.close()
        
            if index%2000 == 0:
                print("index of ok image: ", index)

            index += 1
        print("success for allfiles")
    except TypeError:
        print("%s image is broken. "%file_)

应用举例：

is_valid_file("train2017")

两种方式自行选用，如有问题可以留言～