Rohit8y/Road-Safety-Qwen3-VL-8B-Instruct-bnb-4bit

Model Description

This model is a fine-tuned version of the Qwen3 base model, specialized for road safety understanding and reasoning. It has been instruction-tuned on a custom dataset of 33,000 image-text conversations, where each example consists of an input image and a text prompt, and the model generates an appropriate context-aware response related to road safety.

The dataset used for fine-tuning is custom-built and publicly available here: 🔗 https://huggingface.co/datasets/Rohit8y/Road-Safety-Training-Dataset

Model Sources

• —Repository: [Road-Safety-Vision-Language-Models]