TIGER-AI-Lab/VLM2Vec

View on GitHub

This repo contains the code for "VLM2Vec / MMEB" [ICLR 2025], "VLM2Vec-V2 / MMEB-V2" [TMLR 2026], and "MMEB-V3" [COLM 2026]

Stars: 684
Forks: 64
Open beginner issues: 0
Indexed issues: 31
Avg merge: 8m
Merged PRs (30d): 1

Dominant language: Python
License: Apache-2.0
Last GitHub push: Aug 23, 2026
Latest indexed: Sep 20, 2026
Contributing guide: No contributing guide
Code of conduct: No code of conduct
Beginner labels: No beginner labels indexed

Beginner friendly Everything

Open All

31 open issues indexed

momentseeker数据不对 Open

Difficulty 4/5 3-5 days Newbie friendliness 15/100

TIGER-AI-Lab/VLM2Vec#212 · Opened Sep 8, 2026
video.yaml配置不对 Open

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

TIGER-AI-Lab/VLM2Vec#211 · Opened Sep 8, 2026
memory.yaml配置文件 Open

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

TIGER-AI-Lab/VLM2Vec#210 · Opened Sep 8, 2026
v3 training data Open

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

TIGER-AI-Lab/VLM2Vec#203 · Opened May 17, 2026
评估支持使用vllm框架部署的模型吗 Open

Difficulty 4/5 3-5 days Newbie friendliness 30/100

TIGER-AI-Lab/VLM2Vec#200 · Opened Apr 3, 2026
VD-Vidore-V2 evaluation metrics do not match the paper. Open

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

TIGER-AI-Lab/VLM2Vec#194 · 2 comments · Opened Mar 11, 2026
How to support eval Qwen2.5/3-omni Open

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

TIGER-AI-Lab/VLM2Vec#193 · Opened Mar 5, 2026
one gpu oom Open

Difficulty 4/5 3-5 days Newbie friendliness 25/100

TIGER-AI-Lab/VLM2Vec#192 · Opened Feb 11, 2026
Does this code support training baseline models (GME, LamRA) on private datasets? Open

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

TIGER-AI-Lab/VLM2Vec#191 · Opened Jan 19, 2026
how to eval a openapi model Open

Difficulty 4/5 3-5 days Newbie friendliness 25/100

TIGER-AI-Lab/VLM2Vec#189 · Opened Jan 16, 2026
esg_reports_human_labeled_v2: discrepancy between test corpus size and the number of uploaded images Open

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

TIGER-AI-Lab/VLM2Vec#188 · 1 comment · Opened Jan 16, 2026
why not directly use the original Qwen implementation on huggingface? Open

Replied

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

TIGER-AI-Lab/VLM2Vec#183 · 3 comments · 4 reactions · Opened Dec 20, 2025
Backward is missing in the training_step method of the GradCacheLateProcessTrainer Open

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

TIGER-AI-Lab/VLM2Vec#181 · Opened Dec 10, 2025
Question on CLIP finetuning in VLM2Vec-v1 Open

Difficulty 4/5 3-5 days Newbie friendliness 25/100

TIGER-AI-Lab/VLM2Vec#178 · Opened Nov 26, 2025
gme-Qwen2-VL-2B-Instruct eval error Open

bug evaluation/benchmark

Difficulty 3/5 1-2 days Newbie friendliness 30/100

TIGER-AI-Lab/VLM2Vec#173 · Opened Nov 11, 2025
Reproducibility Issue Open

Difficulty 4/5 3-5 days Newbie friendliness 25/100

TIGER-AI-Lab/VLM2Vec#170 · 1 comment · 2 reactions · Opened Nov 7, 2025
Support for Qwen3 and Modeling file Changes Open

documentation enhancement

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

TIGER-AI-Lab/VLM2Vec#169 · 1 comment · 1 reaction · Opened Nov 7, 2025
Train with mmeb in image_only mode, raise error about picke Open

Replied

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

TIGER-AI-Lab/VLM2Vec#162 · 1 comment · Opened Oct 22, 2025
More Retriever Implementations for Latest Multi-Vector Models Open

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

TIGER-AI-Lab/VLM2Vec#161 · 2 comments · 1 reaction · Opened Oct 22, 2025
Will you update the code repo that supports training with multi-node DeepSpeed? Open

Replied

Difficulty 4/5 3-5 days Newbie friendliness 25/100

TIGER-AI-Lab/VLM2Vec#152 · 1 comment · Opened Sep 30, 2025
how to get pos_clip_name Open

Difficulty 3/5 1-2 days Newbie friendliness 25/100

TIGER-AI-Lab/VLM2Vec#151 · 1 comment · Opened Sep 23, 2025
Hard Negative training code release Open

Replied

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

TIGER-AI-Lab/VLM2Vec#150 · 1 comment · Opened Sep 11, 2025
Significant Performance Gap for the visdoc tasks in VLM2Vec-v2 Qwen2-VL-2B Reproduction Open

Difficulty 4/5 3-5 days Newbie friendliness 25/100

TIGER-AI-Lab/VLM2Vec#149 · 9 comments · Opened Sep 9, 2025
Why don’t we see the special tokens <|vision_start|> and <|vision_end|> in the input? Open

Difficulty 4/5 3-5 days Newbie friendliness 25/100

TIGER-AI-Lab/VLM2Vec#147 · 4 comments · 1 reaction · Opened Aug 28, 2025
Discrepancy in Prompt Construction for LamRA Inference Open

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

TIGER-AI-Lab/VLM2Vec#144 · 1 comment · Opened Aug 19, 2025
datasets.table.CastError: Couldn't cast Open

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

TIGER-AI-Lab/VLM2Vec#143 · 8 comments · Opened Aug 19, 2025
Batch inference: The representation results of a sample repeated k times and n times are different？ Open

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

TIGER-AI-Lab/VLM2Vec#124 · 2 comments · Opened Jul 16, 2025
How to resolve `AssertionError: image_resolution is conflicting with image_decay_factor`? Open

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

TIGER-AI-Lab/VLM2Vec#110 · 6 comments · Opened Jun 20, 2025
How do I add a new vlm? Open

Difficulty 2/5 1-3 hours Newbie friendliness 25/100

TIGER-AI-Lab/VLM2Vec#97 · 7 comments · Opened Jun 15, 2025
Qwen2.5VL Open

Difficulty 4/5 3-5 days Newbie friendliness 25/100

TIGER-AI-Lab/VLM2Vec#66 · 22 comments · Opened Apr 5, 2025
clip,sigplip, blip how to evaluate Open

Difficulty 3/5 1-2 days Newbie friendliness 30/100

TIGER-AI-Lab/VLM2Vec#31 · 5 comments · Opened Jan 21, 2025

TIGER-AI-Lab/VLM2Vec

Get new issues in your inbox