Distributed AI Model Training and LLM Fine-Tuning on Kubernetes
kubeflow/trainer is accelerating — star growth is speeding up by +0.7/day (69th percentile in the tracked cohort), on top of +2.0 stars/day. A recent release is helping fuel the run.
Reconstructed from 199 snapshots — the time-series GitHub’s API doesn’t expose.
Release count rose to 62.
~10 forks/day, 10.1× the trailing baseline.
Release count rose to 61.
Live score, refreshed every 6 hours. Links back to this page.
[](https://repoclimb.com/repo/kubeflow/trainer)Moderate breakout odds over the next 14 days, led by star acceleration (69th pctl) and push recency (99th pctl). Confidence is high given the available history.
Transparent heuristic · logged for model training
Percentile rank within the tracked cohort. The score self-calibrates — it answers “accelerating vs. everything else,” not raw size.