PythonListed by innotive

transformers

huggingface/transformers

164,446 stars34,364 forks

🤗 Transformers: the model-definition framework for state-of-the-art machine learning models in text, vision, audio, and multimodal models, for both inference a

🤗 Transformers: the model-definition framework for state-of-the-art machine learning models in text, vision, audio, and multimodal models, for both inference and training.

audiodeep-learningdeepseekgemmaglmhacktoberfestllmmachine-learningmodel-hubnatural-language-processingnlppretrained-modelspythonpytorchpytorch-transformersqwenspeech-recognitiontransformervlm

Project page

This project has not published an innotive.md page yet.

Repo-native content, AI-friendly workflow

Create innotive.md at the repository root. innotive renders that file as the project page — we do not host an editor.

Keep the content in your repo so it stays versioned with your code. Many teams already run AI in their repositories; drafting innotive.md with your usual AI workflow is usually the fastest path.

Create innotive.md on GitHub

# transformers

One-line summary of what this project does.

## What it solves

Explain the problem and who it's for.

## Quick start

```bash
# install / run commands
```

## Key features

- Feature one
- Feature two

Latest release

Most recent version synced from GitHub.

Patch release: v5.15.1

v5.15.1

Aug 19, 2026

# Patch release v5.15.1 This patch most notably solves a few issues with DFlash and MTP candidate generators, as well as an issue where images could sometimes not be processed on accelerator if using Lanczos filter. It contains the following commits: - Fix DFlash candidate token device mismatch with device_map="auto" (#47877) by @sywangyi and @Cyrilvallez - Align logit distributions for CandidateGenerators using sampling (#48007) by @Cyrilvallez - Fix MTP config when mlp_layer_types is absent (#48015) by @Cyrilvallez - Fallback from 'lanczos' to 'bicubic' when on cuda (#48026) by @zucchini-nlp - Fix gemma4 video to device (#47896) by @guarin