Skip to content

Too many questions in this code #3

@ccyniubi

Description

@ccyniubi

Hello!

In avclassify_model.py -- class BaseClassifierModel(nn.Module):

    def encoder_process(self, modality_data : torch.Tensor, modality_name: str) -> torch.Tensor:
        
        encoder_name = self.enconders[modality_name]['name']
        if encoder_name == 'ResNet18':
            result = self.resnet(modality_data = modality_data, modality = modality_name)
        elif encoder_name == 'Transformer' or encoder_name == 'Transformer_':
            result = self.transformer_process(modality_data = modality_data, modality = modality_name)
        elif encoder_name == 'ViT_B':
            result = self.vit_process(modality_data = modality_data, modality = modality_name)
        elif encoder_name == 'Transformer_LA':
            result = self.transformer_process(modality_data = modality_data, modality = modality_name)

Focus on the " if encoder_name == 'ResNet18': result = self.resnet(modality_data = modality_data, modality = modality_name)" There is even nothing about fuction "self.resnet", there just a fuction named def resnet_process(self, modality_data : torch.Tensor, modality : str) -> torch.Tensor:
Is that fuction right?

Also there so many bugs in this code, I think you may have accidentally uploaded a wrong version. Do you have the latest correct version?
Looking forward to your reply.

Metadata

Metadata

Assignees

No one assigned

    Labels

    No labels
    No labels

    Type

    No type

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions