Open bujiebuhuo opened 1 month ago
This paper aims at zero-shot inference with the original capability of VLMs, which do not need to be trained to be applied to third-party datasets. However, if you pursue further fitting on your dataset, you can try training with a VLM adapter and prompt learning, like FGVP equipped with CoOp.
This paper is great. I want to train this model on my own dataset. Can this code be used to train the model?