bghira/SimpleTuner
A general fine-tuning kit geared toward image/video/audio diffusion models.
What it solves
SimpleTuner 简化了微调大型生成式 AI 模型的过程。它提供一个统一、易用的框架，用于训练图像、视频和音频模型，减少复杂的手动配置和调试，同时支持大量现代模型架构。
How it works
SimpleTuner 充当一个完整的训练流水线，支持多种微调方法，包括 LoRA、LyCORIS 和全秩训练。它集成了 DeepSpeed 与 FSDP2 等高级内存优化工具，使大型模型能够在消费级硬件（最低 16GB VRAM）上训练。项目包含用于生命周期管理的 Web UI 和面向高级用户的命令行界面。还具备嵌入向量自动缓存功能，并与 CaptionFlow 集成进行数据集标注。
Who it’s for
它面向研究人员、AI 艺术家和开发者，让他们在无需深入底层代码的情况下微调生成模型，同时也适用于需要多用户编排、基于角色的访问控制和作业排队的企业团队。
Highlights
- Broad Model Support: Compatible with a massive range of architectures including Flux.1/2, Stable Diffusion XL/3, Wan Video, and LTX Video.
- Multi-Modal Capability: A single pipeline for training image, video, and audio generative models.
- Enterprise-Grade Infrastructure: Includes worker orchestration, SSO integration, and quota management for team-based training.
- Memory Efficiency: Supports quantization (int8/fp8/nf4) and gradient checkpointing to lower hardware barriers.
- Advanced Techniques: Implements TREAD (token-wise dropout), masked loss training, and AnyFlow distillation.