couldn't find pretrained_models/clip-vit-base-patch32-projection

[lluo-Desktop]

Hi， In config : text model name is 'pretrained_models/clip-vit-base-patch32-projection'. But I can't load tokenizer or clip_model by model name. I look into huggingface.co, can't find the model. Can you fix it ?

[xianshunw]

Hi， In config : text model name is 'pretrained_models/clip-vit-base-patch32-projection'. But I can't load tokenizer or clip_model by model name. I look into huggingface.co, can't find the model. Can you fix it ?

same here, any suggestion?

[lluo-Desktop]

Hi， In config : text model name is 'pretrained_models/clip-vit-base-patch32-projection'. But I can't load tokenizer or clip_model by model name. I look into huggingface.co, can't find the model. Can you fix it ?

same here, any suggestion?

I use openAI CLIP_Vit as text encoder : https://huggingface.co/openai/clip-vit-base-patch32 It works.

[wingman-jr-addon]

@wondervictor @lluo-Desktop Any suggestions on an alternate model? I got the above one to work (thanks!), but the license is research-only and non-deployable, so it is more restrictive than the GPL license of this repository posing a barrier to any future deployment.

[onuralpszr]

@wingman-jr-addon maybe you can try https://github.com/mlfoundations/open_clip

[lluo-Desktop]

@onuralpszr @wingman-jr-addon Hi, my friends！Hvae you tired to train yolo-world on objects365/... successfully? I can't not reproduce the model. Any suggestions?

[lluo-Desktop]

I see that the pretrian_model_name has been updated in newest config.