关于训练自己的模型

[susundingkai]

您好，程序卡在data = next(data_iter)，loss等都是nan，请问是python版本问题吗，我用的是3.5

[DayBreak-u]

你用的是什么数据呢，自己的？格式和voc是一样的吗

[susundingkai]

是的，格式和voc是一样的

[DayBreak-u]

修改了读取数据的脚本没？

[susundingkai]

修改了classes，文件夹名保持和voc一致

[DayBreak-u]

那就不太明白了，中间过程debug一下呢？

[susundingkai]

data = next(data_iter)

File "/home/tensor/.local/lib/python3.5/site-packages/torch/utils/data/dataloader.py", line 804, in next idx, data = self._get_data() File "/home/tensor/.local/lib/python3.5/site-packages/torch/utils/data/dataloader.py", line 771, in _get_data success, data = self._try_get_data() File "/home/tensor/.local/lib/python3.5/site-packages/torch/utils/data/dataloader.py", line 724, in _try_get_data data = self.data_queue.get(timeout=timeout) File "/home/tensor/anaconda3/envs/tnet/lib/python3.5/multiprocessing/queues.py", line 104, in get if timeout < 0 or not self._poll(timeout): File "/home/tensor/anaconda3/envs/tnet/lib/python3.5/multiprocessing/connection.py", line 257, in poll return self._poll(timeout) File "/home/tensor/anaconda3/envs/tnet/lib/python3.5/multiprocessing/connection.py", line 414, in _poll r = wait([self], timeout) File "/home/tensor/anaconda3/envs/tnet/lib/python3.5/multiprocessing/connection.py", line 911, in wait ready = selector.select(timeout) File "/home/tensor/anaconda3/envs/tnet/lib/python3.5/selectors.py", line 376, in select fd_event_list = self._poll.poll(timeout)

感觉和python版本有关，开始时next出的数据是对的，后面就卡住了

[DayBreak-u]

可能是的 我是用3.6的

[susundingkai]

感谢回复，问题已经解决了，是pytorch和cv2互锁导致的，关闭cv2的多线程即可。 cv2.setNumThreads(0) cv2.ocl.setUseOpenCL(False)

[246yf87953246]

@susundingkai, 您好，我也出现这个问题，请问这两句代码是添加到哪的？

[ciky-joh]

@susundingkai，方便加好友么，我也问题求教您。q3045696011