train中出现NotImplementedError

blackseven-white commented 2 months ago

训练中出现如下问题，疑似是尺度错误，但不知道是不是环境安装or torch版本的问题，

[0]
===> Start Epoch 1 End Epoch 501
===> Loading datasets
  0%|                                                                                                                         | 0/1000 [00:00<?, ?img/s]
Traceback (most recent call last):
  File "/home/Blackseven/src/ESDNet/train.py", line 169, in <module>
    restored = model_restoration(input_)
  File "/anaconda3/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1501, in _call_impl
    return forward_call(*args, **kwargs)
  File "/home/Blackseven/src/ESDNet/model.py", line 179, in forward
    out_enc_level1 = self.encoder_level1(inp_enc_level1)
  File "/anaconda3/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1501, in _call_impl
    return forward_call(*args, **kwargs)
  File "/anaconda3/lib/python3.10/site-packages/torch/nn/modules/container.py", line 217, in forward
    input = module(input)
  File "/anaconda3/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1501, in _call_impl
    return forward_call(*args, **kwargs)
  File "/home/Blackseven/src/ESDNet/model.py", line 78, in forward
    out = self.residual(x) + self.shortcut(x)
  File "/anaconda3/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1501, in _call_impl
    return forward_call(*args, **kwargs)
  File "/anaconda3/lib/python3.10/site-packages/torch/nn/modules/container.py", line 217, in forward
    input = module(input)
  File "/anaconda3/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1501, in _call_impl
    return forward_call(*args, **kwargs)
  File "/home/Blackseven/.local/lib/python3.10/site-packages/spikingjelly/activation_based/layer.py", line 1832, in forward
    return functional.seq_to_ann_forward(x_seq, super().forward)
  File "/home/Blackseven/.local/lib/python3.10/site-packages/spikingjelly/activation_based/functional.py", line 686, in seq_to_ann_forward
    y = stateless_module(y)
  File "/anaconda3/lib/python3.10/site-packages/torch/nn/modules/batchnorm.py", line 138, in forward
    self._check_input_dim(input)
  File "/anaconda3/lib/python3.10/site-packages/torch/nn/modules/batchnorm.py", line 83, in _check_input_dim
    raise NotImplementedError
NotImplementedError

laofoye99 commented 1 month ago

我遇到了一样的问题，请问有解决么

blackseven-white commented 1 month ago

没有任何人回复我诶🤣还没解决我怀疑可能是cuda版本问题还没验证在 2024年9月4日，18:06，laofoye99 @.***> 写道：我遇到了一样的问题，请问有解决么

—Reply to this email directly, view it on GitHub, or unsubscribe.You are receiving this because you authored the thread.Message ID: @.***>

laofoye99 commented 1 month ago

I have already resolved the issue. The cause of the problem was this line in the model.py file:

layer.ThresholdDendentBatchNorm2d(num_features=dim, alpha=alpha, v_th=v_th, affine=True),

It was mentioned in 请教关于使用tdBN出现的问题 #542 and the solution can be found in 请问框架中的tdBN方法是还没有完成吗？ #391, which requires cloning the Github repository SpikingJelly (version 0.0.0.0.15). However, the latest version available on PyPI is 0.0.0.0.14, which doesn't implement this function, resulting in the error. @MingTian99

MingTian99 / ESDNet

train中出现NotImplementedError #7