All Tools
M
Fine-tuningFreeOpen Source
MS-SWIFT
Fine-tune, align, and deploy LLMs and multimodal models in one framework
Apache-2.0
ABOUT
Teams working with open models often need separate tooling for supervised fine-tuning, alignment, evaluation, quantization, inference, and deployment, which creates duplicated configs and brittle workflows. ms-swift unifies those stages so developers can adapt, evaluate, and serve many text and multimodal models from one framework.
INSTALL
pip install ms-swift -UINTEGRATION GUIDE
1. Fine-tune open LLMs with LoRA or full-parameter training workflows
2. Run RLHF or GRPO-style alignment jobs for better model behavior
3. Train and evaluate multimodal models using text, image, audio, or video inputs
4. Quantize and deploy supported models with backends such as vLLM, SGLang, or LMDeploy
TAGS
pythonfine-tuningmultimodalrlhfgrpoquantizationdeploymentmodelscope