Owner

EvolvingLMMs-Lab

14 indexed repositories · View on GitHub

lmms-eval
One-for-All Multimodal Evaluation Toolkit Across Text, Image, Video, and Audio Tasks

Python · 4413 stars
LLaVA-OneVision-2
Fully Open Framework for Democratized Multimodal Training

Python · 1206 stars
NEO
NEO Series: Native Vision-Language Models from First Principles

Python · 891 stars
lmms-engine
A simple, unified multimodal models training engine. Lean, flexible, and built for hacking at scale.

Python · 825 stars
multimodal-search-r1
[ACL-2026] MMSearch-R1 is an end-to-end RL framework that enables LMMs to perform on-demand, multi-turn search with real-world multimodal...

Python · 486 stars
OneVision-Encoder
Codec-Aligned Sparsity as a Foundational Principle for Multimodal Intelligence

Python · 400 stars
lmms-lab-writer
Agentic LaTeX Writer - Local-first editor for AI-assisted academic writing

TypeScript · 272 stars
LongVT
[CVPR 2026] LongVT: Incentivizing "Thinking with Long Videos" via Native Tool Calling

Python · 266 stars
SkillOpt-Lite
SkillOpt-Lite and HarnessOpt: Optimize your skill or harness with one line of vibe

Python · 187 stars
SimpleStream
A simple video streaming baseline that outperforms SOTAs.

Python · 167 stars
OpenMMReasoner
[CVPR 2026] OpenMMReasoner: Pushing the Frontiers for Multimodal Reasoning with an Open and General Recipe

Python · 166 stars
EASI
Holistic Evaluation of Multimodal LLMs on Spatial Intelligence

Python · 128 stars
ParaVT
ParaVT: Taming the Tool Prior Paradox for Parallel Tool Use in Agentic Video Reinforcement Learning

Python · 58 stars
MacOS-H3-Speedrun
Full MiniMax H3 speedrun workflows for Apple Silicon with strict staging, Turbo, FBC, Spectrum, and SolAttn

Python · 16 stars

Beginner friendly Everything

Open All

VideoMME score is substantially higher than the reported LLaVA-OneVision-7B result with 32 frames Open

bug

Difficulty 4/5 3-5 days Newbie friendliness 50/100

EvolvingLMMs-Lab/lmms-eval#1541 · 4 comments · Opened Sep 19, 2026
GPT6 Astra Open

Difficulty 5/5 Over a week Newbie friendliness 30/100

EvolvingLMMs-Lab/EASI#39 · Opened Sep 12, 2026
mathvision: find_math_answer keeps the FIRST \\boxed answer when a model retracts it and boxes a corrected one Open

Difficulty 2/5 1-3 hours Newbie friendliness 84/100

EvolvingLMMs-Lab/lmms-eval#1511 · 1 comment · Opened Sep 4, 2026
mathvision: find_math_answer deletes the \text command name but keeps its braces — \boxed{\text{Sunday}} is scored wrong against gold Sunday Open

Difficulty 3/5 1-2 days Newbie friendliness 78/100

EvolvingLMMs-Lab/lmms-eval#1509 · 1 comment · Opened Sep 4, 2026
mathvision: _fix_fracs is gated on "sqrt" — \frac12-style shorthand never normalizes, silently failing is_equal for sqrt-free answers Open

Difficulty 2/5 1-3 hours Newbie friendliness 88/100

EvolvingLMMs-Lab/lmms-eval#1507 · 1 comment · Opened Sep 4, 2026
mathvision: _fix_sqrt corrupts indexed roots (\sqrt[3]{8} → \sqrt{[}3]{8}), silently failing the latex2sympy equivalence path Open

Difficulty 2/5 1-3 hours Newbie friendliness 86/100

EvolvingLMMs-Lab/lmms-eval#1505 · 2 comments · Opened Sep 4, 2026
Verify evals on Papers with Code Open

Difficulty 3/5 1-2 days Newbie friendliness 48/100

EvolvingLMMs-Lab/lmms-eval#1483 · Opened Aug 28, 2026
Use `dtype` instead of deprecated `torch_dtype` for transformers >= 4.56 Open

Difficulty 1/5 Under an hour Newbie friendliness 93/100

EvolvingLMMs-Lab/LLaVA-OneVision-2#197 · Opened Aug 28, 2026
minimax_h3_96gb_turbo_4_fast 执行报错 Open

Difficulty 3/5 1-2 days Newbie friendliness 42/100

EvolvingLMMs-Lab/MacOS-H3-Speedrun#1 · 1 comment · Opened Aug 21, 2026
Verify evals on Papers with Code Open

Difficulty 2/5 1-3 hours Newbie friendliness 52/100

EvolvingLMMs-Lab/LongVT#26 · Opened Aug 16, 2026
Verify evals on Papers with Code Open

Difficulty 2/5 1-3 hours Newbie friendliness 62/100

EvolvingLMMs-Lab/ParaVT#4 · Opened Aug 14, 2026
Verify evals on Papers with Code Open

Difficulty 3/5 1-2 days Newbie friendliness 45/100

EvolvingLMMs-Lab/NEO#17 · Opened Aug 14, 2026
[Bug] LiveMath default config expects 2-1-7 split, but downloader and run.sh use 2-2-6 Open

Difficulty 2/5 1-3 hours Newbie friendliness 75/100

EvolvingLMMs-Lab/SkillOpt-Lite#7 · 1 comment · Opened Aug 11, 2026
where is scripts/eval_only.py ? Open

Difficulty 3/5 1-2 days Newbie friendliness 45/100

EvolvingLMMs-Lab/SkillOpt-Lite#6 · 3 reactions · Opened Aug 5, 2026
Text CE loss spikes during visual pre-training Open

Difficulty 5/5 Over a week Newbie friendliness 25/100

EvolvingLMMs-Lab/NEO#16 · Opened Jul 30, 2026
EgoPlan-Bench2 and RoboFAC Open

Difficulty 5/5 Over a week Newbie friendliness 30/100

EvolvingLMMs-Lab/EASI#38 · Opened Jul 27, 2026
About the SFT checkpoint Open

Difficulty 5/5 Over a week Newbie friendliness 25/100

EvolvingLMMs-Lab/ParaVT#3 · Opened Jul 23, 2026
Verify evals on Papers with Code Open

Difficulty 3/5 1-2 days Newbie friendliness 35/100

EvolvingLMMs-Lab/OneVision-Encoder#123 · Opened Jul 23, 2026
Test the fine tune model Open

EvolvingLMMs-Lab/NEO#13 · 4 comments · 1 assignee · Opened Jul 22, 2026
Verify evals on Papers with Code Open

Difficulty 2/5 1-3 hours Newbie friendliness 48/100

EvolvingLMMs-Lab/LLaVA-OneVision-2#193 · Opened Jul 22, 2026
Group-level post-processing hook for custom result aggregation Open

enhancement

Difficulty 5/5 Over a week Newbie friendliness 38/100

EvolvingLMMs-Lab/lmms-eval#1406 · Opened Jul 22, 2026
The >40% rewrite rule introduces undocumented human intervention Open

Difficulty 4/5 3-5 days Newbie friendliness 48/100

EvolvingLMMs-Lab/SkillOpt-Lite#4 · 2 comments · Opened Jul 20, 2026
Do benchmark-specific prompts preload the optimizer with failure diagnoses and solutions? Open

Difficulty 4/5 3-5 days Newbie friendliness 35/100

EvolvingLMMs-Lab/SkillOpt-Lite#3 · 1 comment · Opened Jul 20, 2026
Security and evaluation integrity: no documented isolation boundary for the optimizer coding agent Open

Difficulty 5/5 Over a week Newbie friendliness 35/100

EvolvingLMMs-Lab/SkillOpt-Lite#2 · 1 comment · 1 reaction · Opened Jul 20, 2026
new code has low forward performence in ovobench Open

Difficulty 5/5 Over a week Newbie friendliness 25/100

EvolvingLMMs-Lab/SimpleStream#3 · Opened Jul 19, 2026
`qwen2_5_vl` video metadata fix in #1269 does not affect `second_per_grid_ts`; effective workaround is to pass fps, but fps > 2 can break temporal position ids Open

bug

Difficulty 4/5 3-5 days Newbie friendliness 42/100

EvolvingLMMs-Lab/lmms-eval#1393 · 1 comment · Opened Jul 14, 2026
Opensourced LLM Skill Open

Difficulty 5/5 Over a week Newbie friendliness 25/100

EvolvingLMMs-Lab/SkillOpt-Lite#1 · 2 comments · Opened Jul 10, 2026
Discrepancy between paper's 4-stage training curriculum and the training scripts/data configs in the repo Open

Difficulty 3/5 1-2 days Newbie friendliness 55/100

EvolvingLMMs-Lab/LLaVA-OneVision-2#187 · 2 reactions · Opened Jul 9, 2026
无法复现 LongVT-7B-SFT 在 VideoMME 上的结果 Open

good first issue question

EvolvingLMMs-Lab/LongVT#25 · 2 comments · 1 assignee · Opened Jun 22, 2026
Where can I find the codec-aligned video preprocessing code? Open

Difficulty 5/5 Over a week Newbie friendliness 25/100

EvolvingLMMs-Lab/LLaVA-OneVision-2#184 · 2 comments · Opened Jun 5, 2026
Question about codec preprocessing internals and source availability Open

Difficulty 4/5 3-5 days Newbie friendliness 35/100

EvolvingLMMs-Lab/LLaVA-OneVision-2#182 · 2 comments · Opened Jun 1, 2026
[Data Release] Missing split_180s.jsonl, split_gt10min.jsonl and corresponding video tars in HuggingFace dataset Open

Difficulty 4/5 3-5 days Newbie friendliness 35/100

EvolvingLMMs-Lab/LLaVA-OneVision-2#176 · 1 comment · Opened May 28, 2026
Is there a Thinking/RL variant of LLaVA-OneVision-2? Open

Difficulty 5/5 Over a week Newbie friendliness 20/100

EvolvingLMMs-Lab/LLaVA-OneVision-2#172 · Opened May 28, 2026
Release LLaVA-OneVision-2 artifacts (models, dataset) on Hugging Face Open

Difficulty 4/5 3-5 days Newbie friendliness 35/100

EvolvingLMMs-Lab/LLaVA-OneVision-2#171 · Opened May 27, 2026
Question about fine-tuning LLaVA-OneVision-2 Open

Difficulty 5/5 Over a week Newbie friendliness 25/100

EvolvingLMMs-Lab/LLaVA-OneVision-2#163 · 1 reaction · Opened May 25, 2026
Availability of full fine-tuned VLM weights for inference Open

Difficulty 5/5 Over a week Newbie friendliness 25/100

EvolvingLMMs-Lab/OneVision-Encoder#122 · Opened May 20, 2026
Are there plans to release long video data? Open

Difficulty 5/5 Over a week Newbie friendliness 25/100

EvolvingLMMs-Lab/LLaVA-OneVision-2#146 · 1 comment · Opened May 18, 2026
Missing checkpoint and merge script for LLaVA-OneVision-2-stage-0 (initial unaligned model) Open

Difficulty 4/5 3-5 days Newbie friendliness 35/100

EvolvingLMMs-Lab/LLaVA-OneVision-2#145 · Opened May 18, 2026
OVO-bench evaluation prompts are inconsistent to its official repo. Open

Difficulty 4/5 3-5 days Newbie friendliness 42/100

EvolvingLMMs-Lab/SimpleStream#2 · 1 comment · Opened May 16, 2026
Difference between the paper and the code: global Top-K or local Top-K? Open

Difficulty 2/5 1-3 hours Newbie friendliness 45/100

EvolvingLMMs-Lab/OneVision-Encoder#121 · Opened May 13, 2026
Add Gemma4 and Qwen3.6 support Open

enhancement

Difficulty 4/5 3-5 days Newbie friendliness 45/100

EvolvingLMMs-Lab/lmms-eval#1315 · 1 comment · 1 reaction · Opened Apr 30, 2026
Add llama.cpp support Open

enhancement

Difficulty 5/5 Over a week Newbie friendliness 35/100

EvolvingLMMs-Lab/lmms-eval#1312 · Opened Apr 27, 2026
Hyper-parameters used for reproducibility Open

Difficulty 3/5 1-2 days Newbie friendliness 48/100

EvolvingLMMs-Lab/LLaVA-OneVision-2#109 · 1 comment · Opened Apr 24, 2026
VLM4D (ICCV'25) and R4D-Bench (CVPR'26 Highlight) Open

EvolvingLMMs-Lab/EASI#36 · 1 assignee · Opened Apr 20, 2026
Low GPU utilization and significant bubbles during training Open

Difficulty 4/5 3-5 days Newbie friendliness 35/100

EvolvingLMMs-Lab/OpenMMReasoner#12 · 1 comment · Opened Apr 18, 2026
[Compatibility] LLaVA-OneVision-1.5-8B-Instruct cannot run on transformers>=5.0.0 Open

Difficulty 3/5 1-2 days Newbie friendliness 35/100

EvolvingLMMs-Lab/LLaVA-OneVision-2#108 · 1 comment · Opened Apr 17, 2026
Why is the energy first normalized in each frame before performing top-k selection across all frames? Open

Difficulty 4/5 3-5 days Newbie friendliness 35/100

EvolvingLMMs-Lab/OneVision-Encoder#116 · 1 comment · Opened Apr 15, 2026
qwen3_5是文本图像多模态模型，但目前只能测评文本 Open

bug

Difficulty 3/5 1-2 days Newbie friendliness 55/100

EvolvingLMMs-Lab/lmms-eval#1300 · 3 comments · 1 reaction · Opened Apr 15, 2026
Are all I-frame tokens intended to be preserved in the current implementation? Open

Difficulty 4/5 3-5 days Newbie friendliness 35/100

EvolvingLMMs-Lab/OneVision-Encoder#113 · 1 comment · Opened Apr 14, 2026
Clarification on Intra-frame (I-frame) interval in data preprocessing Open

Difficulty 3/5 1-2 days Newbie friendliness 35/100

EvolvingLMMs-Lab/OneVision-Encoder#112 · 1 comment · Opened Apr 14, 2026

Showing the newest 100

This page lists what was indexed most recently. Advanced filter has the whole inventory, narrowed by language, difficulty and how long a task takes.

Open advanced filter

EvolvingLMMs-Lab

Showing the newest 100

Get new issues in your inbox