oumi
Easily fine-tune (SFT/RL), evaluate, and deploy Qwen, Gemma, or any open agentic LLM/VLM!
What is oumi?
Easily fine-tune (SFT/RL), evaluate, and deploy Qwen, Gemma, or any open agentic LLM/VLM!
Overview of oumi
What it is
Oumi is a fully open-source platform that streamlines the entire lifecycle of foundation models, from data preparation and training to evaluation and deployment. It supports models from 10M to 400B+ parameters and works with both text and multimodal models.
Key features
- Train and fine-tune models using SFT, LoRA, QLoRA, DPO, GRPO, and more.
- Build agentic, tool-using models with tool-calling SFT, executable tool environments, and reinforcement learning over those environments with verl GRPO.
- Work with text and multimodal models including Llama, Gemma, Qwen, DeepSeek, GLM, Phi, and others.
- Synthesize and curate training data, including multi-turn tool-use conversations, with LLM judges.
- Serve models with vLLM, SGLang, or deploy to dedicated endpoints using
oumi deploy. - Evaluate models across standard benchmarks and with multi-criteria rubric judges.
- Run on laptops, clusters, or clouds (AWS, Azure, GCP, Lambda, Modal, Slurm) with a consistent API.
Use cases
- Fine-tuning open LLMs and VLMs for specific tasks.
- Training agentic models that use tools and environments.
- Data synthesis and curation for training.
- Deploying models to production inference endpoints.
- Running large-scale experiments on cloud clusters.
Getting started
Installation options:
### Basic installation
uv pip install oumi
### With GPU support
uv pip install 'oumi[gpu]'
### Latest development version
uv pip install git+https://github.com/oumi-ai/oumi.git
### Using Docker
docker pull ghcr.io/oumi-ai/oumi:latest
docker run --gpus all -it ghcr.io/oumi-ai/oumi:latest oumi --help
### Quick install script (experimental)
curl -LsSf https://oumi.ai/install.sh | bash
CLI usage examples:
### Training
oumi train -c configs/recipes/smollm/sft/135m/quickstart_train.yaml
### Evaluation
oumi evaluate -c configs/recipes/smollm/evaluation/135m/quickstart_eval.yaml
### Inference
oumi infer -c configs/recipes/smollm/inference/135m_infer.yaml --interactive
License and pricing
License: Apache-2.0. Pricing: open source (free).
Based on the project's public README and repository metadata.