The open source research environment for AI researchers to seamlessly train, evaluate, and scale models from local hardware to GPU clusters.
Terminal score: 0–1000 raw, weighted across 4 dimensions. Public score: 0–10 normalized (shown in the 30-day stars chart above).
