In simple English
Training your own artificial intelligence model often requires complex code and deep technical skill. This toolkit lets you take existing language and vision models and train them on your own data using a single command. It supports hundreds of popular open models and handles everything from basic training to testing and deployment. You can run it on your own computer or server if you have a suitable graphics card. It costs nothing to use under an open-source licence.
What it is
A Python framework (a collection of pre-written code packages that help you build software without starting from scratch) for fine-tuning, aligning, and deploying text and multimodal models. It provides a command-line interface and a web-based control panel to manage the entire training workflow.
What it is good for
Solo builders who want to customize open-source language or vision models for a specific project without writing low-level machine learning code from scratch.
- Supports over six hundred text models and four hundred multimodal models out of the box.
- Includes a web-based user interface so you can run training jobs without using the command line.
- Provides built-in tools for testing, quantization, and high-speed inference deployment.
- Requires an NVIDIA graphics card with significant RAM (such as 9 GB or more) to run training locally.
- Setting up all deep learning dependencies and drivers can be difficult for non-programmers.
- Large language models and training jobs consume substantial electricity and take hours to complete.
Writing custom PyTorch training scripts and manual model deployment code
vLLM
- unslothai/unsloth78k stars
Focuses on fast and memory-efficient fine-tuning for large language models.
- hiyouga/LlamaFactory76k stars
Another popular web-based and command-line framework for fine-tuning large models.
- ggml-org/llama.cpp132k stars
A C and C++ library and command-line tool for running language models locally.
- EleutherAI/lm-evaluation-harness14k stars
Runs standard question sets against a language model and reports the scores.
- mudler/LocalAI50k stars
Serves text, vision, speech, image and video models from one self-hosted service that imitates the paid interfaces.
pip install ms-swift -U