Repolode
Submit
transformerlab

transformerlab/transformerlab-app

Open-source ML research platform to train, fine-tune, evaluate and run models locally or on Slurm and cloud GPU clusters.

View on GitHubWebsite
Stars
5,200
Forks
553
Open issues
36
Last commit
3 days ago

What is it?

Transformer Lab is an open-source machine learning platform that puts model downloading, training, chat, and evaluation into one web interface. Individual users can run it on their own machine: on Apple Silicon Macs, on Linux with NVIDIA or AMD GPUs, and on Windows through WSL2 with an NVIDIA GPU.

The platform supports full fine-tuning, LoRA/QLoRA, and RLHF methods such as DPO, ORPO and SIMPO, with trainers for MLX, CUDA and ROCm. It also covers inference through MLX, vLLM, Ollama and Hugging Face Transformers, format conversion between HuggingFace, GGUF and MLX, and diffusion workflows such as LoRA training for Stable Diffusion and Flux.

For teams, the same interface can submit jobs to Slurm clusters or SkyPilot-managed clouds such as AWS, GCP and Azure, and offers shared experiment tracking, a model registry, and one-click Jupyter, VSCode or SSH sessions on remote nodes.

Key features

  • 01Train and fine-tune models with full fine-tuning, LoRA/QLoRA and DPO/ORPO/SIMPO.
  • 02Run inference through MLX, vLLM, Ollama and Hugging Face Transformers.
  • 03Convert models between HuggingFace, GGUF and MLX formats.
  • 04Run LLM-as-a-judge scoring and LM Evaluation Harness benchmarks.
  • 05Submit jobs to Slurm or SkyPilot clusters from one UI and CLI.
  • 06Extend the platform with a Python plugin system and the lab SDK.

Use cases

  • Fine-tuning and chatting with open LLMs on a single Mac or Linux GPU workstation.
  • Training custom LoRA adaptors for Stable Diffusion or Flux on your own image dataset.
  • Letting a research lab queue training runs across a Slurm cluster or cloud GPUs from one interface.
  • Running standardized benchmarks and LLM-based evaluations on model outputs.

Quick start

  1. 1

    Run the official install script from the README

  2. 2

    Change into the install directory

    cd ~/.transformerlab/src
  3. 3

    Start the app with run.sh, then open http://localhost:8338

    ./run.sh

Things to know

  • Requires Apple Silicon on macOS, an NVIDIA or AMD GPU on Linux, or an NVIDIA GPU via WSL2 on Windows.
  • Licensed under AGPL-3.0, a strong copyleft license that may restrict some commercial use.

In collections

AI image generators & Stable Diffusion toolsRun and train AI on GPUsReinforcement learning frameworks & training environments

Topics

#machine-learning#llm#fine-tuning#diffusion#lora#rlhf#model-evaluation#gpu#diffusion-models#electron

FAQ

What is transformerlab-app?
Open-source ML research platform to train, fine-tune, evaluate and run models locally or on Slurm and cloud GPU clusters. Transformer Lab is an open-source machine learning platform that puts model downloading, training, chat, and evaluation into one web interface. Individual users can run it on their own machine: on Apple Silicon Macs, on Linux with NVIDIA or AMD GPUs, and on Windows through WSL2 with an NVIDIA GPU.
Is transformerlab-app free to use? What license does it have?
transformerlab-app is open source under the AGPL-3.0 license. Check the license terms before using it in commercial products.
Is transformerlab-app still maintained?
Actively maintained. The last commit to transformerlab-app was on Nov 5, 2026.
What can I use transformerlab-app for?
Fine-tuning and chatting with open LLMs on a single Mac or Linux GPU workstation. Training custom LoRA adaptors for Stable Diffusion or Flux on your own image dataset. Letting a research lab queue training runs across a Slurm cluster or cloud GPUs from one interface. Running standardized benchmarks and LLM-based evaluations on model outputs.
How do I install transformerlab-app?
Run the official install script from the README Change into the install directory: cd ~/.transformerlab/src Start the app with run.sh, then open http://localhost:8338: ./run.sh
Where can I get transformerlab-app?
The source code and installation instructions are on GitHub: https://github.com/transformerlab/transformerlab-app

Summary generated by AI from the project's README. GitHub stats refresh daily.

Related projects