Mismatch Between Model Weights and Model Structure in Human 3.6M Benchmarks

[Fngg]

I encountered an issue while working with the model weights downloaded from docs/en/model_zoos/video_benchmarks.md under the Human 3.6M Benchmarks section. Specifically, when loading the weights, such as human_simvp_incepu_cos_ep50.pth, human_simvp_vit_cos_ep50.pth, and human_phydnet_cos_ep50.pth, into the corresponding models, there appears to be a mismatch between the model weights and the model structure. Could you please assist me in resolving this issue?