2U1/Qwen-VL-Series-Finetune

An open-source implementaion for fine-tuning Qwen-VL series by Alibaba Cloud.

What it solves

This project provides a comprehensive toolkit for fine-tuning the Qwen-VL series of vision-language models, including Qwen2-VL, Qwen2.5-VL, Qwen3-VL, and Qwen3.5. It simplifies the process of adapting these models to specific tasks, supporting various training methods from full fine-tuning to parameter-efficient techniques like LoRA and QLoRA.

How it works

The repository contains scripts that leverage HuggingFace and the Liger-Kernel for memory-efficient training. It supports multiple training paradigms, including Supervised Fine-Tuning (SFT), Direct Preference Optimization (DPO), and Group Relative Policy Optimization (GRPO). The system handles mixed-modality datasets (images and videos) and provides specific formatting for reasoning-based models to ensure training matches the model's inference-time chat templates.

Who it’s for

It is designed for AI researchers and developers who want to customize Qwen-VL models for domain-specific visual understanding, video analysis, or reasoning tasks.

Highlights

  • Diverse Training Methods: Supports full fine-tuning, LoRA, QLoRA, DPO, and GRPO.
  • Multimodal Support: Capable of training on single-image, multi-image, and video datasets.
  • Reasoning Optimization: Specialized support for reasoning-trace training for Qwen3-VL-Thinking and Qwen3.5 models.
  • Efficiency Tools: Integrates Liger-Kernel and DeepSpeed (Zero2/Zero3) to reduce memory usage and increase speed.
  • Flexible Freezing: Options to freeze the LLM, vision tower, or projector independently, or unfreeze only the top-k layers.
  • Classification Support: Includes scripts for classification fine-tuning with optional MLP layers.

Related

  • Project
    om-ai-lab/VLM-R1

    VLM-R1 is a framework for training R1-style Large Vision-Language Models using GRPO to improve reasoning and generalization in tasks like visual grounding, math, and object detection.

  • Dispatch
    Vision Language Model Alignment in TRL

    Hugging Face has expanded the TRL library to support advanced alignment methods for Vision Language Models, including MPO, GRPO, and GSPO, alongside native SFT support and vLLM integration.

  • Project
    opendilab/LightRFT

    LightRFT is an open‑source, distributed reinforcement‑learning framework for fine‑tuning large language and multimodal generative models. It unifies training back‑ends (DeepSpeed, FSDP) with rollout engines (SGLang, vLLM), supports a range of RL algorithms (GRPO, REINFORCE++, DAPO, etc.), and works across text, image, video, and audio modalities. The library provides ready‑to‑run examples (e.g., GSM8K with Qwen2.5), extensive docs, and Docker images, making it suitable for researchers and engineers building RL‑based alignment pipelines.

  • Dispatch
    Qwen2.5-VL-32B Release Notes

    Qwen has released Qwen2.5-VL-32B-Instruct, a vision-language model optimized via reinforcement learning for superior mathematical reasoning, fine-grained image understanding, and human-aligned responses.

  • Dispatch
    Hugging Face Vision Language Models 2025 Update

    Hugging Face provides a comprehensive overview of the 2024-2025 evolution of Vision Language Models (VLMs), highlighting trends in any-to-any architectures, reasoning models, and the rise of Vision-Language-Action (VLA) models for robotics.