Repository navigation
feat(recognition): publish reference embeddings and the model pipeline for screenshot recognition - #16
Merged
Conversation
…e for screenshot recognition The bundle now carries a gallery of per-kind reference embeddings, made in the build job by the pinned Member and Snap encoders with ONNX Runtime on the CPU, and a models manifest naming ONNX Runtime Web, the card locator, the encoders and the per-kind field and rank readers with their thresholds. Each card also records its level limit from master data. The runtime job stages the encoder models for the build; publication keeps its create-only order, superset check, read-back and pointer-last rules.
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
The recognition bundle moves to
moenotes.recognition-bundle/2:moenotes.embedding-gallery/1): every Member and Snap card of the enabled regions with its Master identity, regions and level limit, plus per-kind reference embeddings computed in the build job by the pinned encoders with ONNX Runtime on the CPU (float32 unit rows, content-addressed buffers).moenotes.recognition-models/1): ONNX Runtime Web, the card locator, the Member and Snap encoders, field readers and rank readers, each with its crop sizes and acceptance thresholds fromrecognition-runtime.json.Pinned files come from the draft release
recognition-runtime-2(BoxLens v0.2.0 models and ONNX Runtime Web 1.30.0).python -m pytest .github/scripts/test_recognition.pypasses.