I have a TF model (not on HF) and want to share it so that other people can easily access and retrain it. How exactly can it be done in HF? How would Autotrain come into the picture (if at all)?
So far when studying the HF functionality I’ve gotten the following impression:
- If you want to share your model for people to use it for inference – share a checkpoint in a model repo.
- If you want to demo your model capabilities – share a demo application in a space repo.
- If you want to retrain an existing model provided in the Transformers library – use Autotrain for you private project.
Some clarification will be appreciated