Am I using the transfer learning using TIMM correctly

vision
1

Hi,
I am trying to do a binary image classification.
I am using model zoo of timm.
I just want to confirm whether I am doing the transfer learning correctly or not.
Below is my model loading code -

import timm
model_transfer = timm.create_model('convnext_tiny_in22k', pretrained=True,num_classes=2) # noqa

Then passing the model to run on the dataset
Please let me know of any other suggestions also, if any.
Thanks.

2

It looks correct to me.
If it is binary image classification, probably you can pass in_chans=1 to timm.create_model(). You can decide based on your use case.

1 Like
3

Images are RGB, I really don’t think I’ll need to change the in_chans though.