gwonny's picture
Create README.md
ab7af85 verified
metadata
license: cc-by-nc-4.0
language:
  - ko

Base Model:

We made a LLM model with tabtoyou/KoLLaVA-v1.5-Synatra-7b (mistral)

Model Description

We use LoRA(r=128, alpha=256), lr=2e-5, mm_projector_lr = 0

Train Detail

CCTV image data(w/ BBox) used, and 100 epoch train

Others

We are making Multi-modal LLM model for Kolon !