search

facebookresearch / mae

PyTorch implementation of MAE https//arxiv.org/abs/2111.06377
Other
6.93k stars 1.17k forks source link

Release of MAE decoder #155

Closed ustcwhy closed 1 year ago

ustcwhy commented 1 year ago

Thanks for your great work ! MAE is very effective on many downstream tasks, but I noticed that you only seem to release the encoder part. Do you plan to release the complete model checkpoint in the future, including the decoder part ?

alekhka commented 1 year ago

I think you can find them on the Colab notebook

It has links to these two weights - 

https://dl.fbaipublicfiles.com/mae/visualize/mae_visualize_vit_large.pth
https://dl.fbaipublicfiles.com/mae/visualize/mae_visualize_vit_large_ganloss.pth

I wanted the base model so I replaced "large" with "base" and luckily they have uploaded it too!

https://dl.fbaipublicfiles.com/mae/visualize/mae_visualize_vit_base.pth
ustcwhy commented 1 year ago

I think you can find them on the Colab notebook

It has links to these two weights -

https://dl.fbaipublicfiles.com/mae/visualize/mae_visualize_vit_large.pth
https://dl.fbaipublicfiles.com/mae/visualize/mae_visualize_vit_large_ganloss.pth

I wanted the base model so I replaced "large" with "base" and luckily they have uploaded it too!

https://dl.fbaipublicfiles.com/mae/visualize/mae_visualize_vit_base.pth

Thanks ! I also found the MAE with its decoder in Huggingfacec~