New model runtime error

[Astronaut323]

Hi, I had no problems loading and running the previous model, but the new version of the model is running with the following errors.

Traceback (most recent call last):
File "/home/songxiaonan/ReDimNet-master/test_1N.py", line 181, in
acc, suspected_daibin_records = compare_folders(model, folder1, folder2, output_csv_1N, labeld_csv, device,
File "/home/songxiaonan/ReDimNet-master/test_1N.py", line 56, in compare_folders
features2[file_path] = extract_features(model, preprocess_audio(file_path, device)).flatten()
File "/home/songxiaonan/ReDimNet-master/test_gpu.py", line 87, in extract_features
features = model(audio_tensor.unsqueeze(0))
File "/home/model/public/songxiaonan/anaconda3/envs/redimnet/lib/python3.9/site-packages/torch/nn/modules/module.py", line 1501, in _call_impl
return forward_call(*args, **kwargs)
File "/home/songxiaonan/ReDimNet-master/redimnet/model.py", line 422, in forward
x = self.spec(x)
File "/home/model/public/songxiaonan/anaconda3/envs/redimnet/lib/python3.9/site-packages/torch/nn/modules/module.py", line 1501, in _call_impl
return forward_call(*args, **kwargs)
File "/home/songxiaonan/ReDimNet-master/redimnet/layers/features.py", line 139, in forward
x = self.torchfbank(x)+1e-6
File "/home/model/public/songxiaonan/anaconda3/envs/redimnet/lib/python3.9/site-packages/torch/nn/modules/module.py", line 1501, in _call_impl
return forward_call(*args, **kwargs)
File "/home/model/public/songxiaonan/anaconda3/envs/redimnet/lib/python3.9/site-packages/torch/nn/modules/container.py", line 217, in forward
input = module(input)
File "/home/model/public/songxiaonan/anaconda3/envs/redimnet/lib/python3.9/site-packages/torch/nn/modules/module.py", line 1501, in _call_impl
return forward_call(*args, **kwargs)
File "/home/songxiaonan/ReDimNet-master/redimnet/layers/features.py", line 52, in forward
return F.conv1d(x, self.flipped_filter).squeeze(1)
RuntimeError: Input type (torch.cuda.FloatTensor) and weight type (torch.FloatTensor) should be the same

[vanIvan]

Hi @Astronaut323 , please make sure both model and input are on the same device:

RuntimeError: Input type (torch.cuda.FloatTensor) and weight type (torch.FloatTensor) should be the same

States that input is on GPU, but model is not. You've probably forgotten to do model.cuda() or model.to(device)