Folder containing the tokenizer, the conversion script to an hf model, as well as a config.json, a configuration_evaclip based on the hf CLIP model, and a modeling_evaclip script based on hf CLIP model. Those are needed to use the model with AutoModel/ AutoTokenizer.
Simply running the conversion script with the right paths, setting save=True, and pushing to hub should allow you to create a model on the hub that anyone can then use with Auto classes.
(edit): I made a mistake in the json file. the "name_or_path" should be changed to: QuanSun/EVA02-CLIP-bigE-14-plus_s9B.

@Leyo Thanks sooo much!!! It's really helpful. I'll try to include smaller models soon.

QuanSun changed pull request status to merged
Your need to confirm your account before you can post a new comment.

Sign up or log in to comment