Any plan to share the pre-trained models?

[sbkim052]

Hi, thanks for sharing this wonderful code. Do you have any plans to share the pre-trained models?

[shizhediao]

same question

[dragen1860]

+1

[dwlmt]

Same question. I'm looking for a model that can perform retrieval as in CLIP or ALIGN and also caption images so would be good if the pretrained weights were made available.

[rom1504]

If you want them make them, there are no such pretrained model in existence with this repo

[gabrielsantosrv]

Same question. I'm looking for a model that can perform retrieval as in CLIP or ALIGN and also caption images so would be good if the pretrained weights were made available.

@dwlmt I am looking for the same model :/

[yxding95]

+1, Hope the authors release the pre-trained model, Thanks!

[hpeiyan]

+10086

[willyfh]

+1

[beybars1]

It would be great, if authors share the pre-trained models.

[mexicantexan]

+9999

[Tangnameless]

+1

[Ella2le]

+1

[AnonymousDestroyer]

+99

[adeetyapatel12]

+5555

[xs1997zju]

+1

[rom1504]

I think there is a confusion here :)

This repo is a reproduction repo. The author of the repo does not have any pretrained model.

If you want to reproduce coca models, you can use this repo (and a lot of compute power) to train some

[TjuJianyu]

+999

[yoon28]

+1

[Timotheeee]

it's made by google so obviously they won't be sharing the models

[germanjke]

I guess we can take ViT weights from ViT-Pytorch library, but NLP part of the model contains many Transformer Blocks, so we don't have weights for them. Looks like we can't get fully weights for CoCa right now.

[rom1504]

if someone feels like plugging https://github.com/lucidrains/CoCa-pytorch/blob/main/coca_pytorch/coca_pytorch.py into https://github.com/mlfoundations/open_clip code base, I could quickly train it

[byougert]

+1

[zhangj1an]

+1

[rom1504]

https://twitter.com/gpuccetti92/status/1621874887844315139?t=l-j6Iv_5d9A48d0XPzvUQg&s=19 done by @gpucce