Stars
Foundation Model Training Using Human Demonstrations
Code and Data for "Language Modeling with Editable External Knowledge"
The code for "TokenPacker: Efficient Visual Projector for Multimodal LLM".
Refine high-quality datasets and visual AI models
Code for Finetune like you pretrain: Improved finetuning of zero-shot vision models
An open source implementation of CLIP.
This repository provides code for machine learning algorithms for edge devices developed at Microsoft Research India.