v0.17.1

What's Changed

• fix: take the package, image and chart version from the release tag by @Yash-Chindam in #45

Full Changelog: v0.17.0...v0.17.1

v0.17.0

What's Changed

• feat: roll back a failed model or policy canary without a person by @Yash-Chindam in #44

Full Changelog: v0.16.0...v0.17.0

v0.16.0

What's Changed

• feat: fall back for streamed requests that fail before their first token by @Yash-Chindam in #43

Full Changelog: v0.15.0...v0.16.0

v0.15.0

What's Changed

• feat: route on the input modalities a request requires by @Yash-Chindam in #42

Full Changelog: v0.14.0...v0.15.0

v0.14.0

What's Changed

• feat: authenticate callers with short-lived signed tokens by @Yash-Chindam in #41

Full Changelog: v0.13.0...v0.14.0

v0.13.0

What's Changed

• feat: make every adapter reproducible from a PEFT or Unsloth recipe by @Yash-Chindam in #40

Full Changelog: v0.12.0...v0.13.0

v0.12.0

What's Changed

• feat: scan the release image and model artifacts before release by @Yash-Chindam in #39

Full Changelog: v0.11.0...v0.12.0

v0.11.0

What's Changed

• feat: install the platform with a Helm chart built from the manifests by @Yash-Chindam in #38

Full Changelog: v0.10.0...v0.11.0

v0.10.0

What's Changed

• feat: serve every catalog model from Ray worker pools split by accelerator by @Yash-Chindam in #37

Full Changelog: v0.9.0...v0.10.0

v0.9.0

What's Changed

• feat: record models, adapters, benchmarks and promotions in MLflow by @Yash-Chindam in #36

Full Changelog: v0.8.0...v0.9.0