ValueError: Input contains NaN, infinity or a value too large for dtype('float32').

panovr commented 3 years ago

Traceback (most recent call last):
  File "train.py", line 97, in <module>
    main()
  File "train.py", line 71, in main
    val_log = val(val_loader,net,criterion,device)
  File "D:\Users\zhaoyili\project\VesselSeg-Pytorch\function.py", line 126, in val
    ('val_auc_roc', evaluater.auc_roc())])
  File "D:\Users\zhaoyili\project\VesselSeg-Pytorch\lib\metrics.py", line 44, in auc_roc
    AUC_ROC = roc_auc_score(self.target, self.output)
  File "D:\Users\zhaoyili\.conda\envs\pytorch\lib\site-packages\sklearn\utils\validation.py", line 63, in inner_f
    return f(*args, **kwargs)
  File "D:\Users\zhaoyili\.conda\envs\pytorch\lib\site-packages\sklearn\metrics\_ranking.py", line 524, in roc_auc_score
    y_score = check_array(y_score, ensure_2d=False)
  File "D:\Users\zhaoyili\.conda\envs\pytorch\lib\site-packages\sklearn\utils\validation.py", line 63, in inner_f
    return f(*args, **kwargs)
  File "D:\Users\zhaoyili\.conda\envs\pytorch\lib\site-packages\sklearn\utils\validation.py", line 720, in check_array
    _assert_all_finite(array,
  File "D:\Users\zhaoyili\.conda\envs\pytorch\lib\site-packages\sklearn\utils\validation.py", line 103, in _assert_all_finite
    raise ValueError(
ValueError: Input contains NaN, infinity or a value too large for dtype('float32').

lixiang007666 commented 2 years ago

我也遇到了这个问题你解决了么

SilverBach805 commented 1 year ago

我也遇到了这个问题，你们解决了吗

liqiangde commented 1 year ago

用Denseunet没这个问题，切换Unet模型训练7个epoch出现了这个问题

Zhuxl123 commented 7 months ago

用Denseunet没这个问题，切换Unet模型训练7个epoch出现了这个问题

用unet训练了20epoch出现了这个问题，请问你解决了吗？

lee-zq / VesselSeg-Pytorch

ValueError: Input contains NaN, infinity or a value too large for dtype('float32'). #3