CatalogWantedMirrorsLedgerOperating StatementDocsStudio statusCard system
alt.freeweights.tools / TRL
From: huggingface <github.com/huggingface>Newsgroups: alt.freeweights.toolsSubject:

TRL

Release: v1.18.0 (2026-11-13)Commit: 480dde9License: Apache-2.0

Supervised fine-tuning and preference training for transformers models.

trainingv1.18.0releaseTRLPythonApache-2.0
FactsLicenseApache-2.0LanguagePythonKindtrainingRelease2026-11-13PlatformsLinux, macOS, WindowsLinks to the projectpermissive
$ pip install trl
# use with Freeweights
$ trl sft --model_name_or_path <id> --dataset_name <dataset>
OverviewInstallUse with FreeweightsLicenseProvenance

Overview

Supervised fine-tuning and preference training for transformers models.

Kind
training
License
Apache-2.0
Language
Python
Platforms
Linux, macOS, Windows
Latest release
v1.18.0 (2026-11-13)
Repository
huggingface/trl

Freeweights links to the project and does not host tool binaries.

Install

$ pip install trl
# use with Freeweights
$ trl sft --model_name_or_path <id> --dataset_name <dataset>
$ pip install trl

Checked against the project's README on 2026-11-14. The project's own instructions win when they differ: install docs

Freeweights links to the project and does not host tool binaries.

Use with Freeweights

Fine-tune a catalog model
$ trl sft --model_name_or_path <id> --dataset_name <dataset>

Download the weights from Freeweights first (HF_ENDPOINT or fw.py), then point the tool at the folder.

Catalog models it can fine-tune (33)

YuE2DeepSeek-V4-Flash-0731DeepSeek-V4.1-Flash-1017gemma-4-26B-A4B-itgemma-4-31B-itGLM-4.7-FlashGLM-5.3GLM-5.3-Flashgpt-oss-120bgpt-oss-20bKimi-K3.1-Dflash2-PreviewKimi-K3.1-ProKimi-Linear-48B-A3B-BaseKimi-Linear-48B-A3B-InstructMiMo-V2.6-Distill-Qwen-9BMiMo-V2.6-Flash-RLMiniMax-M3.1-FlashMinistral-3-14B-Reasoning-2512NVIDIA-Nemotron-3.5-Lightning-30B-A3B-BF16Olmo-3-1025-7BOlmo-3.1-32B-Instructphi-4Qwen-AgentWorld-35B-A3BQwen3-TTS-12Hz-1.7B-BaseQwen3-VL-8B-InstructQwen3.5-4B-BaseQwen3.5-9BQwen3.5-9B-BaseQwen3.6-27BQwen3.6-35B-A3BQwen3.8-27BQwen3.8-Flash-NextUnlimited-OCR-2

License

License
Apache-2.0
Class
Permissive

Read the license

Freeweights links to the project and does not host tool binaries.

Provenance

From: huggingface <github.com/huggingface>Newsgroups: alt.freeweights.toolsSubject: TRLRelease: v1.18.0 (2026-11-13)Commit: 480dde9License: Apache-2.0
Repository
huggingface/trl
Pinned release
v1.18.0, 2026-11-13
Release commit
480dde905c9137898819b27b6c2c6f9be9f34046
Default branch
main 8ad18ec1b6493b85cf9fffbd87d376767ceb8611
Fetched
2026-11-14
Takedowns
None.

Freeweights links to the project and does not host tool binaries.

Get it
InstallProject repository
Works with

33 catalog models. See the list