efficientnetv2_rw_t training input size

[hyxcl]

in previous release we see that efficientnetv2_rw_t weights get top-1 82.34 @ 288x288 .
is 288*288 the training input size or just validation input size?

[rwightman]

@hyxcl
validation (aka 'test' input size in the pretrained cfgs) https://github.com/huggingface/pytorch-image-models/blob/main/timm/models/efficientnet.py#L1044

am going to add eval for both input size in future results tables (for the models that have an additional test res)

[hyxcl]

got it , thank you very much @rwightman