This project aims to simplify the development process of foundation models by providing an end-to-end solution from data preparation to training, evaluation, and deployment. Developed in Python, it allows model training, fine-tuning, evaluation, and deployment through YAML files. It supports models ranging from small-scale (10M parameters) to extremely large-scale (405B parameters), covering text, vision, and multimodal models. Oumi is compatible with various mainstream model architectures (such as Llama series, DeepSeek, Qwen, Phi, etc.) and includes the vLLM efficient inference engine and optimized fine-tuning methods (such as LoRA, QLoRA, DPO).