MLOps Platform
9.2k 2026-04-30
oumi-ai/oumi
An end-to-end platform for fine-tuning, evaluating, and deploying open-source Large Language Models (LLMs) and Vision Language Models (VLMs).
Core Features
End-to-end LLM/VLM lifecycle management (fine-tuning, evaluation, deployment).
Support for various open-source models (gpt-oss, Qwen3, DeepSeek-R1, etc.).
Advanced capabilities like data synthesis, hyperparameter tuning, and model quantization.
Integration with popular ML libraries (Transformers, TRL, vLLM, DeepSpeed).
Deployment to dedicated inference endpoints (e.g., fireworks.ai, parasail).
Quick Start
pip install oumiDetailed Introduction
Oumi is an open-source MLOps platform designed to streamline the entire lifecycle of developing and deploying state-of-the-art foundation models, including Large Language Models (LLMs) and Vision Language Models (VLMs). It provides comprehensive tools for data preparation, fine-tuning, evaluation, and deployment, supporting a wide array of open-source models and integrating with leading ML frameworks. Oumi aims to empower developers and researchers to build custom, high-performance AI models efficiently, from experimentation to production-ready inference endpoints.