AI Model Training

H2O EvalGPT EvalGPT

A large model evaluation system based on the Elo rating method launched by H2O.ai

JAX JAX

A differentiable programming framework from Google designed for high-performance ML research

Cerebras Inference Cerebras

Ultra-fast large model inference API and dedicated endpoint services for developers and enterprises

Cog (Replicate) Cog

An open source containerization tool produced by Replicate that packages ML models into standardized Docker images

Kimi Linear Moonshot AI

Dark Side of the Moon Open Source Hybrid Linear Attention Long Context Model

Cosmos 3 NVIDIA

An open Physical AI world base model that integrates visual reasoning, world generation and action generation into the same model stack

LITELLM LiteLLM

Open source AI Gateway for 100+ LLMs, unifying OpenAI formats, routing, budgeting and observations

DreamGen NVIDIA

NVIDIA Robot Learning Research Project Using Video World Models to Generate Synthetic Training Data

LLaMA-Factory LLaMA-Factory

An integrated open source framework for multi-model fine-tuning and evaluation, covering 100+ LLM/VLM efficient training processes.

FastGPT FastGPT

A knowledge base and visual AI workflow platform based on large models, focusing on the construction of safe and controllable enterprise-level AI agents.

MeteoRA NJU DeepEngine

Nanjing University’s open source multi-task LoRA embedding framework

MLflow MLflow

Open source AI engineering platform covering Agent/LLM observability to traditional ML lifecycle management