v0.9.1
What's Changed
- feat(install): show suggested commands for installer script by @EricLBuehler in #2315
- feat(uqff): no dummy run for uqff generation pass by @EricLBuehler in #2316
- fix(sampler): return natural-log probabilities from
logprobs, not log10 by @dx2102 in #2327 - fix(core): address recent correctness regressions by @EricLBuehler in #2345
- chore(deps): bump serde_with from 3.16.1 to 3.21.0 by @dependabot[bot] in #2349
- feat(lora): add true lora kernels and dynamic adapter support by @EricLBuehler in #2351
- feat(core): support lmhead/embedding quantization by @EricLBuehler in #2280
- fix(quant): misc fixes, warmup metal kernels by @EricLBuehler in #2353
- fix(gemma4): improve audios ordering around text by @EricLBuehler in #2355
- feat(core): major improvements to scheduler & concurrent serving performance by @EricLBuehler in #2354
- feat(core): first-class support for loading GGUFs for all supported models by @EricLBuehler in #2365
- feat(models): support Muse Glimmer by @EricLBuehler in #2366
New Contributors
Full Changelog: v0.9.0...v0.9.1