search

Alibaba-MIIL / ML_Decoder

Official PyTorch implementation of "ML-Decoder: Scalable and Versatile Classification Head" (2021)
MIT License
317 stars 53 forks source link

can you share the backbone pretrain weight on open image ? #17

Closed ckvic3 closed 2 years ago

ckvic3 commented 2 years ago

Thank you for your work! It's a very interesting work. I'm reproducing your's work and trying to do some research based on your work. But the Tresnet-L is too big to train, so I want to start with Tresnet-m. But I am not able to find the pre-trained weights of Tresnet-m on open image dataset. So I'm asking here if you can share the pre-trained weights of the Tresnet series of models on open-image?

mrT23 commented 2 years ago

sorry, i cant at this moment.

you can still use the ImageNet21K pretrain weights, they are also very good https://github.com/Alibaba-MIIL/ImageNet21K

ckvic3 commented 2 years ago

ok, thank you for your reply!