search

wyn430 / CPAC-CNN

MIT License
8 stars 3 forks source link

ValueError: expected sequence of length 8 at dim 1 (got 3) #1

Open yeshihai opened 1 year ago

yeshihai commented 1 year ago

There are some problems appeared when I run the code "python train_mnist.py 8 3 6 10 cuda:0",: # python train_mnist.py 8 3 6 10 cuda:0Traceback (most recent call last): File "train_mnist.py", line 128, in <module> Net = CNN_Decomp(num_filters, filter_h, filter_w, image_channels, rank, devices, num_class, input_shape) File "/home/yeshihai/code/glp/CPAC-CNN/CPAC-CNN-GPU/cnn_torch.py", line 20, in __init__ self.conv1 = Conv_Decomp(num_filters, filter_h, filter_w, image_channels, rank, devices) File "/home/yeshihai/code/glp/CPAC-CNN/CPAC-CNN-GPU/conv_decomp_torch.py", line 31, in __init__ self.factor1 = nn.Parameter(torch.tensor(factors[1]).to(devices[0]))ValueError: expected sequence of length 8 at dim 1 (got 3)

I'm not sure if the problem is caused by the environment or the code itself

wyn430 commented 1 year ago

There are some problems appeared when I run the code "python train_mnist.py 8 3 6 10 cuda:0",: # python train_mnist.py 8 3 6 10 cuda:0Traceback (most recent call last): File "train_mnist.py", line 128, in <module> Net = CNN_Decomp(num_filters, filter_h, filter_w, image_channels, rank, devices, num_class, input_shape) File "/home/yeshihai/code/glp/CPAC-CNN/CPAC-CNN-GPU/cnn_torch.py", line 20, in __init__ self.conv1 = Conv_Decomp(num_filters, filter_h, filter_w, image_channels, rank, devices) File "/home/yeshihai/code/glp/CPAC-CNN/CPAC-CNN-GPU/conv_decomp_torch.py", line 31, in __init__ self.factor1 = nn.Parameter(torch.tensor(factors[1]).to(devices[0]))ValueError: expected sequence of length 8 at dim 1 (got 3)

I'm not sure if the problem is caused by the environment or the code itself

The code is tested on the environment listed in the repo. Could you please double-check the version of "tensorly" and its output? It looks to me because the CP-decomposition results are not as expected.