MCG-NJU / MixFormer

[CVPR 2022 Oral & TPAMI 2024] MixFormer: End-to-End Tracking with Iterative Mixed Attention
https://arxiv.org/abs/2203.11082
MIT License
461 stars 75 forks source link

Performance on GOT-10k #93

Open qkdkralsgh opened 1 year ago

qkdkralsgh commented 1 year ago

Hello, first of all thank you for the good work.

I have one question: Is the performance of the MixViT_L model on the got10k dataset the performance learned on the full dataset? (AO : 75.7%)

yutaocui commented 1 year ago

Hi, the MixViT-L model, which obtains the AO of 75.7% on got10k-test, is trained only on the got10 dataset.

qkdkralsgh commented 1 year ago

Thank you for answer. Is the backbone a convmae large model?

yutaocui commented 1 year ago

Hi, the MixViT-L model, which obtains the AO of 75.7% on got10k-test, is trained only on the got10 dataset.

This model employs ViT-L as backbone. (the MixViT_L(ConvMAE) uses convmae backbone.)

NoorTahir521 commented 10 months ago

The model was trained on GOT 10k full the AO comes to be 57% with lower IOU what can be the reason? it is different than the paper?