【报错】RuntimeError: Unable to find a valid cuDNN algorithm to run convolution

 报错展示：

Traceback (most recent call last):
  File "train_seg_luad.py", line 291, in <module>
    train(cfg=cfg)
  File "train_seg_luad.py", line 179, in train
    _, segs, attns = wetr(inputs)
  File "/home/pengzhang/anaconda3/envs/TPRO/lib/python3.8/site-packages/torch/nn/modules/module.py", line 1110, in _call_impl
    return forward_call(*input, **kwargs)
  File "/data5/pengzhang/TPRO-main/seg_network/model.py", line 51, in forward
    seg = self.decoder(_x)
  File "/home/pengzhang/anaconda3/envs/TPRO/lib/python3.8/site-packages/torch/nn/modules/module.py", line 1110, in _call_impl
    return forward_call(*input, **kwargs)
  File "/data5/pengzhang/TPRO-main/seg_network/segformer_head.py", line 135, in forward
    x = self.linear_pred(x)
  File "/home/pengzhang/anaconda3/envs/TPRO/lib/python3.8/site-packages/torch/nn/modules/module.py", line 1110, in _call_impl
    return forward_call(*input, **kwargs)
  File "/home/pengzhang/anaconda3/envs/TPRO/lib/python3.8/site-packages/torch/nn/modules/conv.py", line 447, in forward
    return self._conv_forward(input, self.weight, self.bias)
  File "/home/pengzhang/anaconda3/envs/TPRO/lib/python3.8/site-packages/torch/nn/modules/conv.py", line 443, in _conv_forward
    return F.conv2d(input, weight, bias, self.stride,
RuntimeError: Unable to find a valid cuDNN algorithm to run convolution

解决方法：

这个错误的原因是咱们的GPU存储不足了。我的程序需要6G的GPU空间，一开始我使用的是显卡0，显卡0还剩余4G左右的空间，显然不够我的程序需要，所以就出现了上面的错误，在我换到显卡3（显卡3还剩余10G左右）以后，程序就可以正常执行了。