TimeMaster-SFT-Qwen2.5-VL-3B-CTU

<h1 align="center">TimeMaster: Training Time-Series Multimodal LLMs to Reason via Reinforcement Learning</h1> <p align="center"> <a href="https://arxiv.org/abs/2506.13705"> <img src="https://img.shields.io/badge/arXiv-Paper-red?style=flat-square&logo=arxiv" alt="arXiv Paper"></a> &nbsp; <a href="https://github.com/langfengQ/TimeMaster"> <img src="https://img.shields.io/badge/GitHub-Project-181717?style=flat-square&logo=github" alt="GitHub Project"></a> &nbsp; <a href="https://huggingface.co/collections/langfeng01/timemaster-68554b6ec27ee539d07a6e40"> <img src="https://img.shields.io/badge/HuggingFace-Models-yellow?style=flat-square&logo=huggingface" alt="HuggingFace Models"></a> </p>

TimeMaster is a reinforcement‑learning‑enhanced framework for training time‑Series multimodal large language models (MLLMs). It enables structured, interpretable reasoning over visualized time‑series signals and has been evaluated on real‑world tasks such as EMG, ECG and Human Activity Recognition (HAR) using Qwen2.5‑VL‑3B‑Instruct.

How To Use

This model is ONLY used as the cold-start SFT model for RL post training of TimeMaster on the CTU dataset. To use this model, please refer to TimeMaster's README.md.