Repository navigation
Conversation
![github-actions[bot]](https://avatars.githubusercontent.com/in/15368?s=60&v=4){kind=small}
Contributor
There was a problem hiding this comment.
go2 replay realtime (arm64)
Details
| Benchmark suite | Current: e156b18 | Previous: 97344c8 | Ratio |
|---|---|---|---|
peak memory |
1728.785 MB |
1602.453 MB |
1.08 |
peak threads |
371 threads |
356 threads |
1.04 |
network (transport) |
2724.566 MB |
2724.755 MB |
1.00 |
disk write |
1.816 MB |
1.742 MB |
1.04 |
instructions |
133.116 G |
114.909 G |
1.16 |
This comment was automatically generated by workflow using github-action-benchmark.
❌ 4 Tests Failed:
View the top 3 failed test(s) by shortest run time
To view more test analytics, go to the Test Analytics Dashboard |
This branch was successfully deployed
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
1 participant
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Detection image, calibration, TF, point-cloud and vision-message consumers now use generated values from
dimos.messages. Local detection collections keep their existing methods; their export methods produce standard generated vision messages for CDR encoding. This removes the legacy rich-message dependency from the migrated YOLO, 2D/3D module and experimental object accumulation paths.Project: https://linear.app/dimensional/project/message-encoding-refactor-05a3ddb4af3e/overview
Stack
Base: #4658 (Python TF and its unchanged #4645 recording prerequisite). Remaining detector families, other perception consumers and the final default transport/legacy retirement are follow-up layers; this draft is not the completed runtime cutover. Final stack bases/descriptions will be aligned after those remaining implementation layers.
File changes
Total relative to #4658: 37 files, 1,308 additions, 532 deletions. Generated-at-build classes are not checked in. The real CDR fixture is a selected conversion of DimOS's approved recording, not copied upstream message definitions.
Manual test
From this PR checkout with the normal perception development dependencies:
For all cases in those two locations, including the fast codec/VLM cases:
uv run pytest dimos/perception/detection/type dimos/models/vl/test_base.py -o addopts='' --import-mode=importlibRun the complete Python block under Perception / Detections for actual CPU YOLO detection, generated message export and cleanup. The first use obtains the existing project sample/model assets. No robot or API key is required. VLM tests mock only the external OpenAI client and exercise actual resize/JPEG/request construction.
Verification
The new fixture archive SHA256 is
966a7b4d542a58a7891b7c0fc4951e3ea3975c56f52cd99b0a5f0541198d2d40. Its manifest records the approved original archive OID51a817f2b5664c9e2f2856293db242e030f0edce276e21da0edc2821d947aad2, source entry hashes, recorder timestamps and per-packet CDR hashes/source nanoseconds. Fixture loading verifies packet hashes and exact stamps. The real image/odometry source-clock offset of approximately 58 seconds is preserved; a lookup using the wrong clock is explicitly rejected rather than rewriting data.Wheel SHA256:
0210891196eaf71505691a6b508ac1d104b9f36450054813d834512e6afa2554.No terminal CI result is claimed; CI has not been polled. Benchmark investigation remains paused. M20 proprietary SDK access, default-host multicast, Nix derivation evaluation, external viewer sign-in and physical/human acceptance remain separate unverified gates. There are also remaining implementation scopes described above; this PR does not claim only external gates remain.
The preserved remote autofix head is
e156b18bdcc1ed0ec1e5ff1d5dc5bf1d3b76a433. Its only change from the validated author commit94c8d724removes three inline lint comments indimos/models/vl/base.py; the parsed Python AST is identical. The categorized table above is pinned to the author commit. No terminal CI result is inferred from this metadata update.