JinXins/Awesome-Token-Merge-for-MLLMs

A paper list about Token Merge, Reduce, Resample, Drop for MLLMs.

31
/ 100
Emerging

This is a curated list of research papers and associated codebases focused on optimizing how large multimodal language models (MLLMs) process visual information. It compiles different techniques like 'token merge' to make MLLMs more efficient, taking research papers as input and providing summaries, links to papers, and code as output. The primary users are AI researchers and practitioners working on MLLM development and efficiency.

Use this if you are a researcher or engineer looking for a comprehensive overview of recent advancements in vision token compression methods for multimodal large language models.

Not ideal if you are a non-technical user seeking a ready-to-use application or a general introduction to large language models.

AI-research multimodal-AI LLM-efficiency computer-vision deep-learning-optimization
No Package No Dependents
Maintenance 6 / 25
Adoption 9 / 25
Maturity 16 / 25
Community 0 / 25

How are scores calculated?

Stars

86

Forks

Language

License

MIT

Last pushed

Oct 26, 2025

Commits (30d)

0

Get this data via API

curl "https://pt-edge.onrender.com/api/v1/quality/transformers/JinXins/Awesome-Token-Merge-for-MLLMs"

Open to everyone — 100 requests/day, no key needed. Get a free key for 1,000/day.