lukemelas / PyTorch-Pretrained-ViT

Vision Transformer (ViT) in PyTorch
770 stars 124 forks source link

Assert 'resize_positional_embedding' prevents creating a model with custom parameters. #30

Open tortueTortue opened 1 year ago

tortueTortue commented 1 year ago

image

The variable "resize_positional_embedding" is created at line 136 and is not in the constructor's parameters. I you want to specify your own parameters and you don't specify a name you get "NameError: name 'resize_positional_embedding' is not defined.", because of line 64.