github AlexsJones/llmfit v1.1.11

6 hours ago

1.1.11 (2026-08-25)

Bug Fixes

  • account for hybrid attention in KV cache (#953) (76ad75d)
  • fit: stop flagging pre-quantized models that fit as insufficient (#898) (e1379a9)
  • ollama: stop a size-less family tag claiming a much larger model (#899) (fabb63b)
  • plan: use active MoE params in fallback estimates (#924) (e11c6e1)
  • providers: match every K-quant and I-quant GGUF variant (#873) (2042c3a)

What's Changed

  • chore(deps): bump http from 1.4.2 to 1.5.0 by @dependabot[bot] in #888
  • chore(deps): bump regex from 1.12.4 to 1.13.1 by @dependabot[bot] in #850
  • chore(deps): bump tokio from 1.52.3 to 1.53.1 by @dependabot[bot] in #847
  • chore(deps): bump serde from 1.0.228 to 1.0.229 by @dependabot[bot] in #844
  • chore(deps): bump dorny/paths-filter from 4.0.2 to 4.0.3 by @dependabot[bot] in #890
  • chore(deps): bump docker/login-action from 4.5.1 to 4.6.0 by @dependabot[bot] in #848
  • chore(deps): bump softprops/action-gh-release from 3.0.1 to 3.0.2 by @dependabot[bot] in #843
  • chore(deps): bump actions/checkout from 7.0.0 to 7.0.1 by @dependabot[bot] in #799
  • chore(deps): bump actions/setup-python from 6 to 7 by @dependabot[bot] in #824
  • chore(deps): group dependabot minor and patch updates by @AlexsJones in #894
  • bench: community results for apple-m1-max by @maxsagt in #896
  • fix(providers): match every K-quant and I-quant GGUF variant by @AlexsJones in #873
  • test(hardware): add mixed GPU fixtures by @luojiyin1987 in #882
  • bench: community results for apple-m4-pro by @akhranovsky in #903
  • fix(fit): stop flagging pre-quantized models that fit as insufficient by @ba1lly in #898
  • bench: add NVIDIA GB10 (DGX Spark class) with 122 models by @Akciali in #872
  • fix(ollama): stop a size-less family tag claiming a much larger model by @ba1lly in #899
  • bench: community results for nvidia-geforce-rtx-3090 by @KeplerAeroIT in #908
  • test(hardware): add Windows GPU parser fixtures by @saime428 in #912
  • fix(plan): use active MoE params in fallback estimates by @HsiangNianian in #924
  • bench: community results for apple-m3-pro by @PakapolTaubol in #910
  • bench: community results for intel-arc-graphics-130v-140v-integrated by @Donixon in #918
  • bench: community results for amd-ati by @viluhh in #936
  • bench: community results for amd-radeon-780m-graphics by @camegone in #947
  • chore(deps): bump the cargo-minor-patch group across 1 directory with 5 updates by @dependabot[bot] in #943
  • [Automated] Weekly benchmark cache refresh by @github-actions[bot] in #891
  • bench: community results for nvidia-geforce-rtx-3050-ti-laptop-gpu by @tingfeng347 in #933
  • fix: account for hybrid attention in KV cache by @AlexsJones in #953
  • docs: place --memory before the recommend subcommand in Termux/Android examples by @junnhwan in #937
  • chore(main): release 1.1.11 by @AlexsJones in #907

New Contributors

Full Changelog: v1.1.10...v1.1.11

Don't miss a new llmfit release

NewReleases is sending notifications on new releases.