EvolvingLMMs-Lab/LLaVA-OneVision-2

Fully Open Framework for Democratized Multimodal Training

About the project

LLaVA-OneVision-2 is an openly released multimodal AI model and training framework that processes images, long-form video, and spatial information. Researchers use it to train, evaluate, and reproduce vision-language models with the project’s released data, encoders, checkpoints, and training records. The catalogue skills support work with this model and its training resources.

1.2kStars on the repository

8Mods indexed here, across every type

5d agoLast push, which is what freshness is scored on

Apache-2.0Licence, which decides whether bodies are shown

Sort Relevance Recently updated Fewest tokens Security grade Repo stars Stars this week Top rated 8 mods

commit-message

EvolvingLMMs-Lab/LLaVA-OneVision-2

Skill Claude CodeCodex

Guide for writing clear, consistent git commit messages following this repository's conventions.

not rated 1.2k 5d ago A 17 tokens original Apache-2.0

cu-lengths-attention-flow

EvolvingLMMs-Lab/LLaVA-OneVision-2

Skill Claude CodeCodex

Bilingual guide for understanding how culengths controls attention behavior across ViT and LLM stages, and how patchpositions scope differs between the two.

not rated 1.2k 5d ago A 35 tokens original Apache-2.0

distributed-offline-packing

EvolvingLMMs-Lab/LLaVA-OneVision-2

Skill Claude CodeCodex

Bilingual guide for running offlinepacking/autopipe.sh across multiple nodes to produce padding-free packed WebDataset shards for SFT, with Energon Metadataset assembly.

not rated 1.2k 5d ago C 40 tokens original Apache-2.0

length-pool-sort-dataset

EvolvingLMMs-Lab/LLaVA-OneVision-2

Skill Claude CodeCodex

Bilingual guide for understanding LengthPoolSortDataset cross-rank length synchronization mechanism in multi-GPU training.

not rated 1.2k 5d ago A 27 tokens original Apache-2.0

llava-onevision2-consistency

EvolvingLMMs-Lab/LLaVA-OneVision-2

Skill Claude CodeCodex

Bilingual guide for running and interpreting LLaVA-OneVision2 HF vs Megatron consistency checks across TP and PP settings.

not rated 1.2k 5d ago A 32 tokens original Apache-2.0

megatron-checkpoint-layout

EvolvingLMMs-Lab/LLaVA-OneVision-2

Skill Claude CodeCodex

Bilingual guidance for Megatron checkpoint 1D 2D 3D mprank layouts across tensor pipeline and expert parallel dimensions.

not rated 1.2k 5d ago A 33 tokens original Apache-2.0

merge-ov2

EvolvingLMMs-Lab/LLaVA-OneVision-2

Skill Claude CodeCodex

Bilingual guide for merging ViT + LLM into LlavaOnevision2 HF checkpoint and validating weight/inference consistency.

not rated 1.2k 5d ago A 29 tokens original Apache-2.0

offline-packing-env-vars

EvolvingLMMs-Lab/LLaVA-OneVision-2

Skill Claude CodeCodex

Bilingual guide for the OFFLINEPACKINGBMR and OFFLINEPACKEDDATA environment variables that control LLaVA-OneVision2 training-side packing — what each gate does, why both must be enabled together, MBS=1 requirement, and the dead OFFLINEPACKINGVQA branch.

not rated 1.2k 5d ago A 65 tokens original Apache-2.0

At most 3 mods per repository are shown here, and a mod shipped inside a plugin is left to that plugin's page — the rest are on their repository pages: