Hugging Face Transformers is a model-definition framework for machine learning models across various modalities. It supports both inference and training, with a focus on ease of use and integration across frameworks.
Collecting history — the radar snapshots this repo daily. The trend line appears after 3 days of data (1 so far).
What it is
Hugging Face Transformers is a model-definition framework for state-of-the-art machine learning models in text, vision, audio, and multimodal models, for both inference and training.
How it works
Transformers centralizes model definitions, ensuring compatibility across various training frameworks (e.g., Axolotl, Unsloth, DeepSpeed, PyTorch-Lightning) and inference engines (e.g., vLLM, SGLang). It supports over 1M+ model checkpoints available on the Hugging Face Hub.
Getting started
Transformers requires Python 3.10+ and PyTorch 2.4+. To install it, create and activate a virtual environment:
# venv
python -m venv .my-env
source .my-env/bin/activate
# uv
uv venv .my-env
source .my-env/bin/activate
Then install Transformers:
# pip
pip install "transformers[torch]"
# uv
uv pip install "transformers[torch]"
To install from source:
git clone https://github.com/huggingface/transformers.git
cd transformers
# pip
pip install '.[torch]'
# uv
uv pip install '.[torch]'
Recent releases
- v5.13.1 Patch release v5.13.1 (2026-07-11): Focused on enabling
transformersfor the latest release of vllm. - v5.13.0 Release v5.13.0 (2026-07-03).
- v5.12.1 Patch release v5.12.1 (2026-06-15): Updated the lower bound for PEFT and fixed auto tokenizer.
- v5.10.4 Patch release v5.10.4 (2026-06-15).
- v5.12.0 Release v5.12.0 (2026-06-12).
Traction
- Stars: 162509
- Forks: 33868
- Open issues: 2480
Behind the repo
The repository is maintained by Hugging Face, a company focused on democratizing AI and machine learning.
Caveats
- License: Apache-2.0
- Open issues: 2480
- Created on: 2018-10-29
- Last push: 2026-07-11.