L'électron rare
3af51d39ed
- Add rewrite_v2_nemo.txt: less directive on scene closure, natural endings - Add repair_v2_nemo.txt: same approach, avoids mechanical conclusion formulas - PromptStore.render() now accepts prompt_profile kwarg with fallback to v1 - Config: [prompt_profiles] maps model to profile (e.g. mistral-nemo -> v2_nemo) - Mark qwen2.5:7b as accepted baseline Ollama via Mascarade P2P - Close 3 P0 TODOs Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
88 lines
2.6 KiB
TOML
88 lines
2.6 KiB
TOML
[paths]
|
|
mascarade_repo = "/Users/electron/Documents/Projets/mascarade"
|
|
# Route tout via Mascarade core local — Mistral/Apple/Ollama sont des providers
|
|
core_base_url = "http://127.0.0.1:8100"
|
|
apple_runtime_url = "http://127.0.0.1:8201"
|
|
ollama_tags_url = "http://127.0.0.1:11434/api/tags"
|
|
# `openai_compatible` envoie `ollama:*` via Mascarade /v1/chat/completions
|
|
# qui route vers le provider Ollama (P2P vers VM) ou Mistral selon le prefixe.
|
|
ollama_runtime = "openai_compatible"
|
|
ollama_openai_base_url = "http://127.0.0.1:8100"
|
|
apple_model_ready_timeout_seconds = 30
|
|
apple_model_poll_interval_seconds = 2
|
|
|
|
[smoke]
|
|
chapter = "02"
|
|
intention = "Chapitre court. Une femme arrive dans une ville de nuit, trouve un indice simple, et finit sur une decision risquee. Style direct, phrases courtes, ton sobre."
|
|
timeout_seconds = 300
|
|
|
|
[preset]
|
|
ANE_PROVIDER = "openai_compatible"
|
|
ANE_BASE_URL = "http://127.0.0.1:8100" # Mascarade core local — route vers Mistral/Apple/Ollama
|
|
# ANE_API_KEY doit venir de l'environnement shell (export ANE_API_KEY=...)
|
|
ANE_TIMEOUT = "300"
|
|
ANE_MAX_TOKENS_STRUCTURE = "256"
|
|
ANE_MAX_TOKENS_DRAFT = "768"
|
|
ANE_MAX_TOKENS_CRITIQUE = "512"
|
|
ANE_MAX_TOKENS_REWRITE = "1024"
|
|
ANE_MAX_TOKENS_GATE = "384"
|
|
ANE_MAX_TOKENS_REPAIR = "1536"
|
|
ANE_MAX_TOKENS_MEMORY = "320"
|
|
ANE_REPAIR_MAX_PASSES = "2"
|
|
|
|
[ensure_models]
|
|
apple_models = [
|
|
"qwen3-4b-instruct-2507-q4f16",
|
|
"qwen2.5-0.5b-instruct-onnx",
|
|
"qwen3.5-4b-onnx-q4f16",
|
|
]
|
|
ollama_models = [
|
|
"qwen2.5:7b",
|
|
"qwen2.5:1.5b",
|
|
"mistral-nemo:latest",
|
|
]
|
|
|
|
[lots.priority_models]
|
|
models = [
|
|
"mistral:mistral-large-latest",
|
|
"ollama:qwen2.5:7b",
|
|
"apple-coreml:qwen3-4b-instruct-2507-q4f16",
|
|
]
|
|
|
|
[lots.baselines]
|
|
models = [
|
|
"apple-coreml:qwen2.5-0.5b-instruct-onnx",
|
|
"ollama:qwen2.5:1.5b",
|
|
]
|
|
|
|
[lots.french_models]
|
|
models = [
|
|
"mistral:mistral-large-latest",
|
|
"ollama:mistral-nemo:latest",
|
|
"apple-coreml:qwen3-4b-instruct-2507-q4f16",
|
|
]
|
|
|
|
# Prompt profiles per model (optional — falls back to v1 if not set)
|
|
[prompt_profiles]
|
|
"ollama:mistral-nemo:latest" = "v2_nemo"
|
|
|
|
[lots.preflight_only]
|
|
models = []
|
|
|
|
[tracking.ane]
|
|
todo_active = "TODO_ACTIVE.md"
|
|
todo_done = "TODO_IMPLEMENTE.md"
|
|
plan = "docs/EXECUTION_PLAN_2026-03-22.md"
|
|
comparison = "docs/MODEL_COMPARISON_2026-03-22.md"
|
|
readme = "README.md"
|
|
runbook = "docs/runbooks/LOCAL_GENERATION.md"
|
|
|
|
[tracking.mascarade]
|
|
todo = "TODO_AI_NOVEL_ENGINE.md"
|
|
plan = "docs/EXECUTION_PLAN_2026-03-08.md"
|
|
readme = "README.md"
|
|
runbook = "docs/RUNBOOK_APPLE_LLM_LOCAL.md"
|
|
|
|
[next_actions]
|
|
rewrite_compaction = "Rejouer priority_models avec mistral:mistral-large-latest comme reference qualite + qwen3-4b-instruct-2507 comme ref locale."
|