Oumi
Apache-2.0 end-to-end foundation-model platform for training, evaluation, inference and agentic/tool-use post-training, installable as PyPI oumi 0.9.
Tags
Description
Established project (repo oumi-ai/oumi; not archived; created 2024-05-07; not a 2026 launch). Root LICENSE is Apache-2.0. PyPI oumi 0.9 verified 2026-11-12 (not yanked; wheel uploaded 2026-09-16; project URLs point at this repository). README install: uv pip install oumi (optional oumi[gpu]), or pip; a git install is also documented. Docs at https://oumi.ai/docs/en/latest/index.html (HTTP 200). README news for v0.9 describes an end-to-end stack for agentic, tool-using models, including tool-calling SFT/DPO, executable tool environments, and GRPO over Oumi environments with verl. Executable environments can run tool or task code. GPU, cloud job runners (README mentions GCP, AWS, Azure, Lambda) and model-weight licenses are separate; no cloud accounts were used. Vendor: Oumi.
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.
- verl: Distributed reinforcement-learning training library for LLMs, with GRPO/PPO dataflows and pluggable training and rollout engines for agentic post-training.
- 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.
Sources
- https://oumi.ai
- https://github.com/oumi-ai/oumi
- https://github.com/oumi-ai/oumi/blob/main/LICENSE
- https://github.com/oumi-ai/oumi/blob/main/README.md
- https://pypi.org/project/oumi/
- https://oumi.ai/docs/en/latest/index.html
Machine-readable
- JSON
- Markdown
- Source file on GitHub (edit via pull request)