Select remote utterance embedding model

This commit is contained in:
Meta Vens 2026-05-19 17:09:20 +01:00
commit fb79c85250
2 changed files with 8 additions and 5 deletions

1
.gitignore vendored
View file

@ -6,6 +6,7 @@ state/*.msgpack
state/*.msgpack.lock
state/*.projection.json
state/scratch-compaction-*.md
state/scratch-runtime/
.weksa-runs/
.weksa-smoke/
node_modules/

View file

@ -43,9 +43,9 @@ constraints:
pronunciations, phrase grouping, stress, and broad IPA stringing.
- Long-term speech handoff should use learned line-expression vectors from
Weksa and learned synth-control decoding in AquaSynth.
- Large offline text-embedding jobs should use the neighboring Ollama
inference box documented in `E:\Projects\gamecult-ops` rather than running
on the local Weksa workspace by default.
- Large offline text-embedding jobs should use `bge-m3` on the neighboring
Ollama inference box documented in `E:\Projects\gamecult-ops` rather than
running on the local Weksa workspace by default.
invariants:
- Ontology owns meaning.
@ -186,8 +186,8 @@ open_questions:
morphology, and diachrony definitions?
- What compiled runtime bundle format keeps game-time rendering fast and
inspectable?
- Which Ollama embedding model on `192.168.1.84` should be the default for
offline semantic decomposition and critique?
- How should Weksa persist or cache utterance embeddings generated by
`bge-m3` on the neighboring Ollama box?
- How much deterministic validation belongs in v0?
- What is the minimal Rust to Dust language-project slice that proves the
engine without hard-coding the setting into it?
@ -224,3 +224,5 @@ landed_baseline:
- First English pronunciation plan fixture for Nibu proof output.
- First toy line-expression vector fixture for Nibu proof output.
- First Nibu interlingua-to-English lowering fixture.
- `bge-m3:latest` installed on neighboring Ollama box `192.168.1.84` and
verified through a 1024-dimensional utterance embedding.