- stage-d: next-line / block-close / multi-line fixes, symbols format 2, TS errors in the header.
Held-out: 0.426 exact / 0.700 edit sim (stage-c 0.405 / 0.689); block close 0.27 -> 0.79.
- stage-e: next-edit (80% exact region rewrites, abstains on 100% of no-edit samples).
- Manifest: nextEditModel (companion) and nextEditFormat (older extensions ignore it, so they
never use next-edit with the empty-output deletion bug).
- Daemon: next-edit served by the companion model in a second llama-server, started after the
main model loads.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01J39imRDT8L7xYbWzsVWd8A
llama-server leaves special tokens out of content, so the NO_EDIT token arrives as an
empty string. The daemon treated that as a rewrite of the region to nothing (Tab would
have deleted the region). shared/nextedit.ts parseNextEditOutput() is now used by the
daemon and the eval.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01J39imRDT8L7xYbWzsVWd8A
Parses Qwen-format FIM prompts (prefix- or suffix-first, optional repo/file headers) back into
prefix/suffix and rebuilds them in the trained layout; mode from the model string (ac:symbols /
ac:simple); files located via workspaceRoots by path or content. extensions/zed/README.md documents
the Zed settings; untested against real Zed traffic yet.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01J39imRDT8L7xYbWzsVWd8A
- shared/nextedit.ts: next-edit prompt format (region rewrite or NO_EDIT, Qwen spare special tokens)
- shared/diagnostics.ts: TS errors near the cursor as context; symbols format 2 adds namespace imports
- shared/prompt.ts: token healing (cut the half-typed word, force it via grammar), errors header section
- daemon: model profiles carry contextFormat/errors/members/nextEdit; /v1/next-edit with edit-burst
tracking; /v1/prewarm; /v1/debug/last; token healing on by default
- extension: next-edit decorations (Tab apply / Esc dismiss), suggest on cursor move, prewarm on focus,
Show Last Prompt, update check against releases, diagnostics in the edit recorder
- models/README: training-data attribution; release script profiles for stage-d/stage-e
- .gitea/workflows/ci.yml: build + test, publish release on vscode-v* tags (needs a runner + RELEASE_TOKEN)
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01J39imRDT8L7xYbWzsVWd8A
- models/: GGUFs (Git LFS), manifest.json, model card, Apache-2.0 LICENSE + NOTICE
- extensions/vscode/: the Local Autocomplete extension (v0.5.0, publisher harvmaster)
- daemon/: acd completion daemon (bundled into the extension; also runs standalone)
- shared/: prompt layout + symbol extraction shared with the training pipeline
- scripts/: make-model-release.mjs, publish-extension-release.mjs (Gitea releases)
- README with quick start, settings, architecture, build and release instructions;
MIT license for code; CHANGELOG
The default model manifest URL moves to models/manifest.json.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
stage-c (default): Qwen2.5-Coder-0.5B fine-tuned for suffix-first FIM with
repo/symbol-context headers. stage-b: same without context headers.
manifest.json lists files, SHA-256s, prompt profiles and the pinned llama.cpp
build for the Local Autocomplete extension's managed mode.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
HarvmasterandClaude Opus 5.5