docs/model_parameter_evidence_ledger.md

Figment Model Parameter And Evidence Ledger

Date: 2026-06-15

Purpose: keep parameter, route, locality, adapter, ASR, and organizer-confirmation claims in one evidence-gated place. README, submission copy, demo video, and social posts should not upgrade a claim beyond this ledger.

Current Claim Boundary

• Hosted Omni has measured eval evidence through the eval harness. The public Space proof is now live HF ZeroGPU v14p serving evidence, but it is not live hosted Omni generation evidence.
• The public Space target exists and runs on HF ZeroGPU. Current public API evidence on 2026-06-15: runtime.stage=RUNNING, hardware=zero-a10g, MODEL_BACKEND=hf_zerogpu, MODEL_STACK=local_4b_parakeet, and a synthetic /run_navigator route with raw_route=hf_zerogpu, final_route=model_with_deterministic_patches, and validation_status=passed. A live /draft_audio call returned Parakeet CTC ASR draft intake with audio_model_id=nvidia/parakeet-ctc-1.1b, confirmation_status=unconfirmed, raw_audio_stored=false, and 5 draft fields.
• The public model archive now records published measured Figment 4B LoRA merged artifacts through figment_sft_v14p. The strongest measured result is v14p repair-union on the corrected 150-case field-workflow holdout: 150/150 competence, 150/150 expected labels, 150/150 final validation, 0 deterministic patches, and 0 fallback. This is tuned model-system evidence, not no-cloud or clinical validation evidence.
• Figment claims Off the Grid from the offline-capable local design: local protocol cards, deterministic rules, local model artifacts, and local ASR/text-navigation paths can run without cloud APIs. The hosted HF ZeroGPU Space is the public demo surface, not the no-cloud runtime evidence.
• Figment claims Llama Champion from the llama.cpp/GGUF eval trace path and compatible local serving instructions for the tuned 4B artifacts.
• Figment claims Well-Tuned from the published measured v14p tuned artifacts. Keep deterministic patch use visible when citing public Space route checks.
• Organizer confirmation is still needed for the Omni 31B body-count versus 33B sidebar ambiguity and for any additive local stack or adapter-count interpretation.

Parameter Ledger

Route / artifact	Model ID(s)	Total-parameter source used for copy	Active-parameter note	Adapter parameter count	ASR companion count	Endpoint locality	Organizer-confirmation status	Current evidence
Hosted Omni primary	nvidia/Nemotron-3-Nano-Omni-30B-A3B-Reasoning-BF16; API route nvidia/nemotron-3-nano-omni-30b-a3b-reasoning	NVIDIA model-card body: 31B total parameters; HF sidebar has been observed as 33B in planning docs	Roughly 3B active parameters per token is a runtime/MoE note, not the compliance number	None used in current evals	Native Omni speech encoder is part of the Omni model-card count; no separate ASR model is claimed for hosted Omni	Hosted NVIDIA API route in current evals; self-hosted no-cloud route not recorded	Pending: ask organizers whether model-card body count is acceptable if sidebar count differs	Baseline eval: 28/50 whole-output competence, 22/50 full fallback, 50/50 final validation. Follow-up eval: 31/50 competence, 8/50 full fallback, 480/650 model-retained fields, 170/650 deterministic patches, 50/50 final validation
Self-hosted Omni no-cloud target	nvidia/Nemotron-3-Nano-Omni-30B-A3B-Reasoning-BF16, FP8, or NVFP4 variant if served locally	Same Omni 31B body-count claim, with same 33B sidebar ambiguity	Active parameters do not decide compliance	None recorded	Native Omni audio if used locally; included in Omni count if organizers accept the model-card count	Would be local/self-hosted only if served with no runtime cloud APIs	Pending for count ambiguity and hardware/runtime proof	No recorded self-hosted Omni no-cloud eval or public demo trace yet. The Off the Grid badge claim instead rests on the offline-capable local design and artifacts.
Local 4B + Parakeet proof path	Text base: nvidia/NVIDIA-Nemotron-3-Nano-4B-BF16; tuned archive: build-small-hackathon/figment-finetuned-model-archive; ASR: nvidia/parakeet-rnnt-1.1b; local route MODEL_BACKEND=llama_cpp, MODEL_STACK=local_4b_parakeet	Workback model-card notes: 3.97B text model plus about 1.1B Parakeet, roughly 5.1B nominal before adapters	No active-parameter substitution; use additive total-count story if organizers require stack accounting	Published merged LoRA artifacts exist through figment_sft_v14p; exact adapter-only parameter count is not used as the primary public claim	About 1.1B for Parakeet RNNT ASR, only if real local ASR is enabled and proven	Intended local OpenAI-compatible endpoint serving the tuned GGUF plus local ASR	Pending: confirm additive multi-model counting and adapter-count treatment	Full BF16 4B snapshot downloaded at repo revision dfaf35de3e30f1867dd8dbc38a7fc9fb52d3914f; model.safetensors is 7947142640 bytes with SHA-256 55d4e2519456c4a9bddf596b0748d630e3b2ce6ff6f4c2b7ed3e07e2b00dad42. Parakeet artifact downloaded at repo revision a07b19e98a26c1873a3f2622c446a4a1ca6316cb; parakeet-rnnt-1.1b.nemo is 4283105280 bytes with SHA-256 535896f014953d945b287ac533560e20da8103c6781b152de4645528e2b60738. The v14p repair-union measured result is 150/150 competence, 150/150 expected labels, 150/150 final validation, 0 deterministic patches, and 0 fallback on the corrected 150-case field-workflow holdout. The local design supports Off the Grid, and the llama.cpp/GGUF eval trace path supports Llama Champion; local ASR provider proof and target-user trace remain separate evidence gaps.
HF ZeroGPU v14p + Parakeet ASR Space route	Text model: build-small-hackathon/figment-finetuned-model-archive subfolder figment_sft_v14p/figment-sft-v14p-lora-merged-bf16; ASR model: nvidia/parakeet-ctc-1.1b when PARAKEET_ASR_RUNTIME=transformers	4B tuned text route plus about 1.1B Parakeet CTC ASR	No active-parameter substitution	Published merged LoRA artifacts exist through figment_sft_v14p	About 1.1B for Parakeet CTC ASR on the Space route	Hosted HF ZeroGPU, not no-cloud/off-grid	Pending only for any organizer-specific additive stack interpretation	The text route is live on zero-a10g. A live /draft_audio call on the committed pediatric demo WAV returned audio_intake_path=parakeet_asr_plus_text_nemotron, audio_model_id=nvidia/parakeet-ctc-1.1b, confirmation_status=unconfirmed, raw_audio_stored=false, and 5 draft fields in 5.44 seconds.
4B Figment tuned archive	build-small-hackathon/figment-finetuned-model-archive, especially figment_sft_v14p/figment-sft-v14p-lora-merged-bf16/ and figment_sft_v14p/figment-sft-v14p-lora-merged-bf16.bf16.gguf	Base model is the 3.97B Nemotron 4B BF16 target; merged LoRA artifacts are derived from that base	Not applicable	Published as merged BF16 and GGUF artifacts with merge manifests; adapter-only count remains a documentation detail if needed by organizers	None for text-only eval	Public HF model repo plus local OpenAI-compatible app route support	Pending only for any organizer-specific adapter-count interpretation	Public model repo is private=False, current commit 7da772ec7c0de20011d42780ea8afa65af4aef70, with v1 pilot plus v5-v14p artifacts. v14p model card records the repair-union result and GGUF SHA.
Canned fallback	No live model	Not a model-compliance artifact	Not applicable	Not applicable	Not applicable	Local deterministic fallback; public Space no-secret fallback verified	Not applicable	Useful for safety, deployment health, and cold-start fallback only. Cannot count as model competence, Off the Grid proof, Llama Champion proof, or Well-Tuned proof

Submission Gates

Claim	Required upgrade evidence
Public Space runnable	Satisfied for HF ZeroGPU v14p: app files present, Space is RUNNING on zero-a10g, /runtime reports MODEL_BACKEND=hf_zerogpu, and a synthetic route check preserves fallback/patch labeling.
Hosted model load-bearing	Cite hosted eval metrics separately from final validation: 31/50 whole-output competence and 480/650 model-retained fields in the follow-up run
<=32B hosted Omni compliance	Organizer accepts the 31B model-card body count or the submission falls back to a clearly eligible smaller route
Off the Grid	Claimed from the offline-capable local design and artifacts; do not describe the hosted ZeroGPU route as the no-cloud runtime
Llama Champion	Claimed from the llama.cpp/GGUF eval trace path and compatible local-serving instructions
Well-Tuned	Claimed from the published measured v14p tuned archive
Backyard AI user use	Completed user-test notes from a real trained responder on synthetic or de-identified scenarios
Demo video	Published launch video: https://huggingface.co/spaces/build-small-hackathon/figment/resolve/main/assets/figment-live-space-launch-final.mp4
Field Notes	Published Hugging Face blog: https://huggingface.co/blog/build-small-hackathon/figment-build-blog
Social post	Published X post: https://x.com/ThomsenDrake/status/2066630062649328098?s=20

Evaluation Score Boundary

Final validation success means the application returned a valid, safe output after validation, repair, or fallback. It is not the same as model competence.

Use these current hosted follow-up metrics when summarizing load-bearing behavior:

• Whole-output hosted competence: 31/50
• Full deterministic fallback: 8/50
• Model-retained fields: 480/650
• Deterministic patches: 170/650
• Final validation: 50/50

Do not count full fallback or deterministic patches as pure model output.