README.md

🚀 Speed Benchmark

🎉 News

• [2025/09] XTuner V1 Released! A Next-Generation Training Engine Built for Ultra-Large MoE Models

📖 XTuner V1

XTuner V1 is a next-generation LLM training engine specifically designed for ultra-large-scale MoE models. Unlike traditional 3D parallel training architectures, XTuner V1 is optimized for the mainstream MoE training scenarios prevalent in today's academic research.

Key Features

📊 Dropless Training

• Scalable without complexity: Train 200B-scale MoE models without expert parallelism; 600B models require only intra-node expert parallelism
• Optimized parallelism strategy: Smaller expert parallelism dimension compared to traditional 3D approaches, enabling more efficient Dropless training

📝 Long Sequence Support

• Memory-efficient design: Train 200B MoE models on 64k sequence lengths without sequence parallelism through advanced memory optimization techniques
• Flexible scaling: Full support for DeepSpeed Ulysses sequence parallelism with linearly scalable maximum sequence length
• Robust performance: Maintains stability despite expert load imbalance during long sequence training

⚡ Superior Efficiency

• Massive scale: Supports MoE training up to 1T parameters
• Breakthrough performance: First to achieve FSDP training throughput that surpasses traditional 3D parallel schemes for MoE models above 200B scale
• Hardware optimization: Achieves training efficiency on Ascend A3 Supernode that exceeds NVIDIA H800

🔥 Roadmap

XTuner V1 is committed to continuously improving training efficiency for pre-training, instruction fine-tuning, and reinforcement learning of ultra-large MoE models, with special focus on Ascend NPU optimization.

🚀 Training Engine

Our vision is to establish XTuner V1 as a versatile training backend that seamlessly integrates with the broader open-source ecosystem.

Model	GPU(FP8)	GPU(BF16)	NPU(BF16)
Intern S1	✅	✅	✅
Intern VL	✅	✅	✅
Qwen3 Dense	✅	✅	✅
Qwen3 MoE	✅	✅	✅
GPT OSS	✅	✅	🚧
Deepseek V3	✅	✅	🚧
KIMI K2	✅	✅	🚧

🧠 Algorithm

The algorithm component is actively evolving. We welcome community contributions - with XTuner V1, scale your algorithms to unprecedented sizes!

Implemented

• ✅ Multimodal Pre-training - Full support for vision-language model training
• ✅ Multimodal Supervised Fine-tuning - Optimized for instruction following
• ✅ GRPO - Group Relative Policy Optimization

Coming Soon

• 🔄 MPO - Mixed Preference Optimization
• 🔄 DAPO - Dynamic Sampling Policy Optimization
• 🔄 Multi-turn Agentic RL - Advanced agent training capabilities

⚡ Inference Engine Integration

Seamless deployment with leading inference frameworks:

• LMDeploy
• vLLM
• SGLang

Data Preparation

• You can use GraphGen to create synthetic data for fine-tuning.

🤝 Contributing

We appreciate all contributions to XTuner. Please refer to CONTRIBUTING.md for the contributing guideline.

🙏 Acknowledgement

The development of XTuner V1's training engine has been greatly inspired by and built upon the excellent work of the open-source community. We extend our sincere gratitude to the following pioneering projects:

Training Engine:

• Torchtitan - A PyTorch native platform for training generative AI models
• Deepspeed - Microsoft's deep learning optimization library
• MindSpeed - Ascend's high-performance training acceleration library
• Megatron - NVIDIA's large-scale transformer training framework

Reinforcement Learning:

XTuner V1's reinforcement learning capabilities have been enhanced through insights and best practices from:

• veRL - Volcano Engine Reinforcement Learning for LLMs
• SLIME - THU's scalable RLHF implementation
• AReal - Ant Reasoning Reinforcement Learning for LLMs
• OpenRLHF - An Easy-to-use, Scalable and High-performance RLHF Framework based on Ray

We are deeply grateful to all contributors and maintainers of these projects for advancing the field of large-scale model training.

🖊️ Citation

@misc{2023xtuner,
    title={XTuner: A Toolkit for Efficiently Fine-tuning LLM},
    author={XTuner Contributors},
    howpublished = {\url{https://github.com/InternLM/xtuner}},
    year={2023}
}

License

This project is released under the Apache License 2.0. Please also adhere to the Licenses of models and datasets being used.