Hugging Face's logo

facebook
/
convnextv2-huge-22k-512

Image Classification
Transformers
PyTorch
convnextv2
vision
Inference Endpoints
Model card Files Community
4

Image_size should be 512

#1
by dinhanhx - opened
Discussion
dinhanhx

Image_size should be 512 because the model can receive 512 x 512 as written in model's repo name and card.

https://huggingface.co/facebook/convnextv2-huge-22k-512/blob/38dac533dd1105e541fa2be1f783bb23146e4fa8/config.json#L1021

adirik

Hi @dinhanhx , thanks for pointing it out and you are right. This doesn't effect anything at the moment but would be an issue in the tensorflow version of the model.

Sign up or log in to comment