Loading all previous models and their scores?

Hello folks,

I'm trying to utilize models with pre-calculated BrainScore or ImageNet accuracy for downstream tasks but have struggled to find adequate documentation.

Specifically, in PyTorch, how can one:
1) Load each of published models, and 2) Retrieve their ImageNet performance and Brain scores?

(So that one can compare brain scores with other downstream task performances of those models)

+ I also noticed that the vision/brainscore_vision/models directory primarily contains CORnets (might not include all previously submitted models?)

Thank you

Gwangsu

Start by inspecting the documentation and the vision/brainscore_vision/models directory mentioned in the issue, then trace how published models are exposed and how their ImageNet and Brain scores are stored. Done means documenting how to load each available model and retrieve both score types, including whether the directory covers all previously submitted models.

Written by the indexing model from the issue text.