YtongXie / UniMiSS-code

[ECCV2022&TPAMI] Official pytorch implementation of UniMiSS & UniMiSS+
MIT License
57 stars 6 forks source link

Finetune for 2D classification task #3

Closed liyi01827 closed 2 weeks ago

liyi01827 commented 1 year ago

Hi, This is an interesting work. Can you also provide the code for fine-tune on 2d image classification tasks? The hyper-params are missing for these downstream 2d tasks. Thanks a lot.

liyi01827 commented 1 year ago

For "fine-tune" in the paper, does it mean freeze the encoder and only train a linear classifier or train them together?

YtongXie commented 1 year ago

ean freeze the encoder and only train a linear classifier or train them

"fine-tune" means training both encoder and linear classifier together

YtongXie commented 1 year ago

Hi, This is an interesting work. Can you also provide the code for fine-tune on 2d image classification tasks? The hyper-params are missing for these downstream 2d tasks. Thanks a lot.

We have provided the code for fine-tuning 3d classification task and will provide the code for fine-tuning 2d tasks later.

coolboy497 commented 11 months ago

Hi, @YtongXie,Your work catch my eyes, when will you provide the code for fine-tuning 2d tasks ? I very interested to see it, thank you very much

YtongXie commented 2 weeks ago

We have provided the code for fine-tuning 2d tasks. https://github.com/YtongXie/UniMiSS-code/tree/main/UniMiSSPlus/Downstream/2D