Compared to the original BLIP code, the following changes have been made by Zihao Yue: deleted: BLIP.gif deleted: CODEOWNERS deleted: CODE_OF_CONDUCT.md deleted: LICENSE.txt deleted: README.md deleted: SECURITY.md deleted: cog.yaml deleted: eval_nocaps.py deleted: eval_retrieval_video.py deleted: models/blip_itm.py deleted: models/blip_nlvr.py deleted: models/blip_pretrain.py deleted: models/blip_retrieval.py deleted: models/nlvr_encoder.py deleted: train_vqa.py deleted: predict.py deleted: pretrain.py deleted: train_nlvr.py deleted: train_retrieval.py modified: train_caption.py modified: configs/caption_coco.yaml modified: demo.ipynb