torchtune
PyTorch-native BSD-3-Clause library for authoring, fine-tuning and evaluating LLMs with native PyTorch training primitives.
Tags
Description
The official README states that torchtune is no longer actively maintained and that development wound down in 2025, linking to "The future of torchtune" (https://github.com/meta-pytorch/torchtune/issues/2883). Chosen schema-valid status deprecated (the schema has active/beta/deprecated/unknown, not archived or unmaintained). The GitHub repository meta-pytorch/torchtune is not archived (GitHub API archived=false, disabled=false as of 2026-11-11). PyPI torchtune 0.6.1 remains the current listed release, is not yanked, is BSD-3-Clause, and is still installable. Established PyTorch project (not a 2026 launch). Docs at https://pytorch.org/torchtune/main/ currently redirect to https://meta-pytorch.org/torchtune/main/. GPU and model-weight requirements are separate.
Agent access
Related
- TRL: Hugging Face Transformer Reinforcement Learning library for post-training methods such as SFT, DPO, PPO and GRPO on top of transformers.
- Axolotl: Apache-2.0 post-training / fine-tuning framework for a wide range of LLMs, with CLI, YAML configs, Docker and distributed-training options.
- LLaMA-Factory: Apache-2.0 toolkit to fine-tune 100+ LLMs and VLMs via zero-code CLI and the LLaMA Board Gradio web UI.
- Unsloth: Open-source LLM fine-tuning library and desktop app for running and training models locally, with a dual Apache-2.0 / AGPL-3.0 license split.
Guides & comparisons
- Fine-tuning frameworks compared — Axolotl, LLaMA-Factory and Unsloth (Comparison): Compare Axolotl, LLaMA-Factory and Unsloth as SFT, LoRA/QLoRA and preference-tuning frameworks for tool-use models: documented methods, config style, chat/tool datasets, hardware backends, licences and current versions. Cross-links the agent RL stacks comparison.
- Post-training a tool-use model — TRL, PEFT, and scale-out RL (Stack): A recommended post-training stack for chat and tool-use models: TRL and PEFT for SFT, LoRA/QLoRA, DPO and GRPO, with OpenRLHF and AReaL as the distributed and asynchronous RL options. Cross-links the fine-tuning frameworks and agent RL stacks comparisons.
Sources
- https://github.com/meta-pytorch/torchtune
- https://github.com/meta-pytorch/torchtune/blob/main/README.md
- https://github.com/meta-pytorch/torchtune/issues/2883
- https://github.com/meta-pytorch/torchtune/blob/main/LICENSE
- https://pypi.org/project/torchtune/
- https://meta-pytorch.org/torchtune/main/
Machine-readable
- JSON
- Markdown
- Source file on GitHub (edit via pull request)