Export the candidate to ONNX and benchmark CPU, GPU, and device targets #33

New Issue

timmy · 2026-08-19T01:59:01Z

timmy commented

2026-08-19 01:59:01 +00:00

Parent epic: #4

Outcome

Replace the VLM hot path with a compact deployable runtime when evidence permits.

Acceptance criteria

Export parity against reference inference is tested.
p50/p95 latency, memory, startup, and throughput are reported.
Timmy schema adapter and fallback are verified end to end.

Dependencies

#32 — Calibrate confidence and define the production abstention policy

Required evidence

Test or executable receipt attached to the issue/PR
Privacy/safety boundary checked where applicable
Clean checkout and relevant full suite pass

Delivery

Open a focused branch and PR with Closes #ISSUE. Do not include real medical images in issue comments or Git history.

Parent epic: #4 ## Outcome Replace the VLM hot path with a compact deployable runtime when evidence permits. ## Acceptance criteria - [ ] Export parity against reference inference is tested. - [ ] p50/p95 latency, memory, startup, and throughput are reported. - [ ] Timmy schema adapter and fallback are verified end to end. ## Dependencies - #32 — Calibrate confidence and define the production abstention policy ## Required evidence - Test or executable receipt attached to the issue/PR - Privacy/safety boundary checked where applicable - Clean checkout and relevant full suite pass ## Delivery Open a focused branch and PR with `Closes #ISSUE`. Do not include real medical images in issue comments or Git history.

timmy added this to the M2 — Consented Dataset & Specialist Model milestone 2026-08-19 01:59:01 +00:00

timmy added the

type/ops

priority/P1

area/model

labels 2026-08-19 01:59:01 +00:00

timmy referenced this issue

2026-08-19 01:59:03 +00:00

EPIC: Specialist Bristol classifier & calibration #4

Sign in to join this conversation.