LaonA2_VL_3B / README.md
gykim
[Init] Models
0f5f063
---
datasets:
- lmms-lab/RefCOCOg
language:
- en
base_model:
- Qwen/Qwen2.5-VL-3B-Instruct
pipeline_tag: zero-shot-object-detection
---
# LaonA2 VL 3B
LaonA2 VL 3BλŠ” Qwen 2.5 VL 3B 기반의 ν–₯μƒλœ λΉ„μ „-μ–Έμ–΄ λͺ¨λΈμž…λ‹ˆλ‹€. VLM-R1 κ°•ν™”ν•™μŠ΅μ„ 톡해 REC(Referring Expression Comprehension) μ„±λŠ₯이 κ°œμ„ λ˜μ—ˆμŠ΅λ‹ˆλ‹€.
cite: arxiv.org/abs/2504.07615