README.md

metadata

license: apache-2.0
datasets:
  - Tuwhy/MIS_Train
base_model:
  - Qwen/Qwen2-VL-7B-Instruct
pipeline_tag: image-text-to-text
tags:
  - safety
  - fine-tuning
  - multi-image
  - mllm

Rethinking Bottlenecks in Safety Fine-Tuning of Vision Language Models

Our paper, code, data, models can be found at MIS.

Description

Qwen2-VL-7B-Instruct model fine-tuned on MIS training set.

MIRgae

Here is example pipeline of MIS training set and MIRage safety CoT label construction.

You can fine-tune Qwen2-VL series using LlamaFactory.