This repository provides full reproduction code for all experiments, figures, and tables in the paper. Given the same datasets and pretrained model weights, running the pipeline will produce results ...

- Stage 1 :: Projection Matrix Alignment between Vision Encoder & Pretrained LLM on CC-3M-595K (Custom) - Stage 2 :: Projection & LLM Finetuning on LLaVa v1.5 Instruct (including various ...