← TinyGPT · docs · devlog · roadmap · speedup
source: docs/pace-model-manifest.md · view on GitHub ↗

Pace model manifest — every model role, decided and on disk

Canonical inventory of all models Pace needs (goal set 2026-06-10: all roles decided with runnable artifacts by end of day). One row per role; this doc is the single source of truth — update it when any decision changes. Embedded/ANE ports are perf infrastructure, not blockers: interim runtime is LM Studio / tinygpt serve, which Pace already uses.

RoleModelArtifact (on this machine)SizeStatus 2026-06-10
Planner (runtime default)qwen/qwen3-30b-a3bLM Studio18.6 GBON DISK, wired (Pace default)
Planner (specialist)v11 DoRA on Qwen3-0.6B, served int8 g64~/.cache/tinygpt/runs/pace-planner-v11/baked-hf~0.5 GB eff.TRAINING — ship-gate verdict tonight; ships only if it beats the gate, else 30B stays
Planner (ANE perf variant)M8 28-block chain, int8 per-block~/.cache/tinygpt/ane-w8b/476 MBON DISK, numerics gate PASS, 17–22 tok/s
Screen VLMA/B tonight: UI-Venus-1.5-2B-6bit vs Qwen3-VL-2B-Instruct-6bit (UI-Venus-8B as third column)all three in LM Studio2.2–5.8 GBDECIDING tonight (scripts/archive/eval_pace_vlm_ab.py, 12 fixtures, FakePace baseline 5/12); M4 Swift port (#266) follows the winner
ASR (voice)WhisperKit openai_whisper-large-v3-turbo~/Documents/huggingface/models/argmaxinc/whisperkit-coreml/openai_whisper-large-v3-v20240930_turbo1.5 GBON DISK, qualified 2026-06-08 (9× realtime on ANE, perfect “Pace” accuracy); app integration = #292 (Swift work, not a model gap)
TTSApple AVSpeechSynthesizer (system)none — OS framework (LocalTTSClient.swift)0DONE — no model needed, zero-cloud by construction
Embeddings (RAG)Qwen3-Embedding-0.6B (decided 2026-06-09, replaced mxbai)~/.cache/tinygpt/models/qwen3-embedding-0.6b/1.2 GBON DISK; runtime retrieval is currently Spotlight-first (PaceSpotlightRetrievalConnector), embedding RAG is a stub contract (PaceLocalRetrieval.swift) — model ready when wiring lands
Dictation post-procStage A rule-based (no model)in pace repo, 26 tests0SHIPPED; Stage B model is conditional on a v11 BFCL trigger, deliberately not built

Disk budget

On-disk total for the shipping set (30B planner + VLM winner + ASR + embeddings): ~22–24 GB, dominated by the 30B. If v11 ships and replaces the 30B for planning, the set drops to ~5–6 GB. The specialist question is therefore also a 4× bundle-size question.

What “done today” means per role

Stale items this manifest supersedes