search

nii-yamagishilab / PartialSpoof

BSD 3-Clause "New" or "Revised" License
33 stars 4 forks source link

Unable to run inference #5

Closed cyaaronk closed 2 months ago

cyaaronk commented 6 months ago

Dear Author,

I try to follow the steps in PartialSpoof/tree/main/03multireso under the section "Using Pretrained Models" when I run : $bash 00_run.sh 2

I encounter error in multi-reso/01/output/log_output_dev: size mismatch for m_transform.0.post_nets_utt.layers.4.proj2.weight: copying a param with shape torch.Size([32, 8]) from checkpoint, the shape in current model is torch.Size([8, 2]). size mismatch for m_transform.0.post_nets_utt.layers.4.proj2.bias: copying a param with shape torch.Size([32]) from checkpoint, the shape in current model is torch.Size([8]). size mismatch for m_transform.0.post_nets_utt.fc.weight: copying a param with shape torch.Size([16, 32]) from checkpoint, the shape in current model is torch.Size([4, 8]).

Why it says the dimensions are different?

zlin0 commented 3 months ago

Sorry, I missed your question earlier. Have you already solved it, or do you still need assistance? Could you provide more information about the issue? Which pre-trained model are you using?

I am guessing you might be using a different SSL model as referenced in model.py:57 and model.py:180 compared to the downloaded one.

zlin0 commented 2 months ago

I will close it, please reopen it if you still have the same issue.