Skip to content

0863: catalog offers the installed engine first; GGUF builds for 39 curated chat rows #13

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
lpalbou merged 4 commits into from
Oct 29, 2026
Merged

0863: catalog offers the installed engine first; GGUF builds for 39 curated chat rows #13

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
4 commits
Select commit Hold shift + click to select a range
4c44c6e
0863: a host's installed engine is offered first; GGUF builds for eve…
lpalbou Oct 29, 2026
5093306
Merge origin/main
lpalbou Oct 29, 2026
626e901
Regenerate llms-full.txt (scripts/update_llms.py)
lpalbou Oct 29, 2026
1f51f31
Merge remote-tracking branch 'origin/main' into fix/catalog-installed…
lpalbou Oct 29, 2026
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
8 changes: 8 additions & 0 deletions CHANGELOG.md
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -40,6 +40,14 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
Defaults grid reads "installed: served by the installed GGUF build ...". Measured there: 1024 dimensions,
similarity 0.775 for two cat sentences vs 0.262 against a revenue sentence. Only without any build the route reads
"not downloaded" with Download.

- **A host's installed engine is offered first in the download catalog** (backlog 0863). On a Linux server with
llama.cpp and neither LM Studio nor Ollama, the Qwen3.5 9B row listed and pre-selected the LM Studio build.
39 curated chat rows had no GGUF build. Each now has one: a Q4_K_M, read from the Hugging Face API on
2026-10-29, with its size and revision. A curated pick for an absent engine yields to an artifact an installed
engine serves. Every artifact carries `engine_installed`, and a row lists the pre-selection first, then
installed engines' artifacts.

- **Structured output with a host's `cancel_event`** no longer fails on the HuggingFace GGUF lane with "got multiple
values for keyword argument 'cancel_event'" (backlog 0861): the structured branch hands the event over once,
under the provider's cancel kwarg.
Expand Down
Loading
Loading