v0.6.6

BREAKING CHANGES

• feat: move /completions under /v1/completions by @tomasmcm in #112

New Contributors

• @tomasmcm made their first contribution in #112

Full Changelog: v0.6.5...v0.6.6

v0.6.5

HOTFIX: correctly resolve initial config

Full Changelog: v0.6.4...v0.6.5

v0.6.4

What's Changed

• feat: Added Orca V2 by @LLukas22 in #65
• feat: UX perf improve for long convo by @louisgv in #73
• chore: sort the model list by @louisgv in #80
• feat: Initialize a shared SDK package, licensed under MIT by @louisgv in #81
• feat: theme switcher by @louisgv in #82

Full Changelog: v0.5.2...v0.6.4

v0.6.3

fix: sdk build

v0.5.2

What's Changed

• feat: Added Nous Hermes to known models by @LLukas22 in #53
• feat: GPU inferencing, cuda for linux/windows/mac intel, metal for mac aarch64 by @louisgv in #56
• fix: Inference on thread view will stop when swapping between route
• fix: Properly terminate prompt feeding when stream stopped
• feat: Inference status text/status comment
• fix: add CUDA setup for linux and windows by @louisgv in #59
• fix: disable gpu toggle if no GPU is available by @louisgv in #63

NOTE: GPU inferencing is only available to Mac Metal (M1/M2) ATM, see #61

local.ai.GPU.Toggle.mp4

New Contributors

• @LLukas22 made their first contribution in #53

Full Changelog: v0.4.0...v0.5.2

v0.5.1

🥷

ref: #61

v0.4.0

What's Changed

• feat: Model Info popup by @Regular-Baf in #36
• feat: improve thread UX, fix linux dev run by @louisgv in #37
• feat: prompt template, persistent thread config side-panel by @louisgv in #38
• feat: save last used model for each thread by @louisgv
• feat: remote tokenizer by @louisgv
• feat: model integrity check now apply default settings if known by @louisgv
• docs: Add git submodule handling to dev workflow by @nightscape in #43
• fix: model stats fallback by @louisgv in #46

New Contributors

• @nightscape made their first contribution in #43

Full Changelog: v0.3.0...v0.4.0

v0.3.0

What's Changed

• feat: WizardCoder's hashes and citations by @Regular-Baf in #32
• feat: model config, with custom remote tokenizer by @louisgv in #35
• fix: directory opening in Linux by @Regular-Baf in #33
• fix: WizardCoder size by @Regular-Baf in #34

New Contributors

• @Regular-Baf made their first contribution in #32

Full Changelog: v0.2.9...v0.3.0

v0.2.9

What's changed:

• Added experimental model tag
• Enable downloader to skip hash check for experimental model

Full Changelog: v0.2.8...v0.2.9

v0.2.8

What's Changed

• feat: quick model switcher by @louisgv in #24
• feat: start server from quick model selector by @louisgv in #25
• feat: type-safe invoke implementation by @louisgv in #26
• feat: Adding completion params passing to API by @louisgv in #27
• chore: cleanup ux, prepare for prompt params tweak by @louisgv in #28

Full Changelog: v0.2.7...v0.2.8