5 Commits

Include renames

Author	SHA1	Message	Date
xavierk	731b919865	chore(release): prepare v0.19.0 (#53) ... Version 0.19.0 across version.py, the Void template and the synced docs; changelog entry; Gitea project URLs in pyproject.toml; per-backend ladder thresholds measured in #19; the xbps release workflow is manual-only, and packaging/void/publish-release.sh publishes a locally built, verified repository to the Gitea releases.	2026-11-06 20:23:14 +00:00
xavierk	2a76793109	fix(app): detect CUDA outside PATH, warn on IBus pin on niri, query Gitea for updates, repoint migrated dictionary (#36) ... CUDA toolkit detection also checks CUDA_HOME, CUDA_PATH and /usr/local/cuda*, and the backend order names only backends the installed build compiled in. An explicit IBus pin on niri is honoured with one warning. The update checker queries this fork's Gitea releases and packaged installs make no network request. The one-time legacy import repoints a migrated dictionary path at the copied file.	2026-11-06 14:38:43 +00:00
xavierk	180cd4d512	fix(spec16): close the review findings for the Void port (#34) ... Model recommendation keyed to the installed build and free memory, fork-sourced Void templates, signed xbps release workflow on Gitea with the pinned repository key, niri injection guard and wtype-first order, tightened naming sweep, web copy fixes, and CI security and duplication fixes.	2026-11-06 14:13:49 +00:00
soubarnak	987d8eec29	fix(models): use ticket 19 measured peaks for ladder thresholds ... q8_0 threshold becomes 1644 MiB (1132+512) and q5_0 becomes 1358 MiB (846+512) per the recorded baseline recommendation, replacing the on-disk-size proxies. Tests pin the measured values and the 1644/1358 boundaries.	2026-11-06 09:00:47 +05:30
soubarnak	df771d8fb1	feat(models): memory-aware whisper.cpp ladder with conditional backend order ... Free memory picks the load-time rung (large-v3-turbo-q8_0 default, q5_0 under 1086 MiB free, small/small.en floor) for unpinned default configs; thresholds derive from catalog sizes plus the protocol's 512 MiB headroom. A manual pin is never overridden, and a failed load steps one rung down and retries exactly once before surfacing. CUDA ranks before Vulkan only with a complete architecture-matching toolkit (nvidia-smi GPU, nvcc, compute capability >= 7.0); otherwise Vulkan, then CPU. The driver-only CUDA recommendation now demotes to CPU logic instead of naming an unreachable VRAM ladder. Refs #20	2026-11-06 01:48:20 +05:30