Skip to content

Add YOLO26 depth export support - #233

Merged
rolandocortez merged 2 commits into
mainfrom
feat/yolov26-depth-map-output
Sep 8, 2026
Merged

rolandocortez merged 2 commits into
mainfrom
feat/yolov26-depth-map-output

Conversation

@rolandocortez

@rolandocortez rolandocortez commented Sep 7, 2026

Copy link
Copy Markdown
Contributor

Summary

Adds YOLO26-Depth support to tools.

  • Detects YOLO26 depth models and routes them through the YOLO26 exporter.
  • Updates the pinned Ultralytics revision to include official YOLO26-Depth support.
  • Exports the dense depth output as output_yolo26.
  • Generates an NNArchive using MapOutputParser with min_max_scaling=False.
  • Adds the model to the E2E test matrix and validates the generated NNArchive contract.

Validation

  • Exported YOLO26n-Depth to ONNX and NNArchive successfully.
  • Validated the generated MapOutputParser NNArchive contract.
  • Ran the model successfully on:
    • RVC2
    • RVC4 FP16
    • RVC4 INT8
  • Compared the converted outputs against the PyTorch YOLO26-Depth model on the exact same decoded frames:
    • RVC2 mean MAE: 0.0500
    • RVC4 FP16 mean MAE: 0.0258
    • RVC4 INT8 mean MAE: 0.1439
  • Measured isolated NN throughput:
    • RVC2: 8.46 FPS
    • RVC4 FP16: 56.85 FPS
  • YOLO26 detect / segmentation / pose / semantic-segmentation regression tests
  • YOLOv8 detect / segmentation / pose / OBB / classification regression tests.

Summary by CodeRabbit

  • New Features

    • Added support for converting YOLO26 depth models.
    • Added depth-model detection and NNArchive generation with map output support.
    • Added validation for depth-model output selection and archive configuration.
  • Tests

    • Added end-to-end coverage for depth-model conversion and output configuration.
    • Updated test sharding assignments to include the expanded test suite.

@coderabbitai

coderabbitai Bot commented Sep 7, 2026

Copy link
Copy Markdown

Review Change Stack

Important

Review skipped

Review was skipped due to path filters

⛔ Files ignored due to path filters (1)
  • media/coverage_badge.svg is excluded by !**/*.svg

CodeRabbit blocks several paths by default. You can override this behavior by explicitly including those paths in the path filters. For example, including **/dist/** will override the default block on the dist directory, by removing the pattern from both the lists.

⚙️ Run configuration

Configuration used: defaults

Review profile: CHILL

Plan: Team

Run ID: 9075a08d-2b8e-43c1-8460-ffd6b2adf8ad

You can disable this status message by setting the reviews.review_status to false in the CodeRabbit configuration file.

Use the checkbox below for a quick retry:

  • 🔍 Trigger review

No actionable comments were generated in the recent review. 🎉

ℹ️ Recent review info
⚙️ Run configuration

Configuration used: defaults

Review profile: CHILL

Plan: Team

Run ID: 69e9d751-15f2-463b-967d-6a52a0e5f463

📥 Commits

Reviewing files that changed from the base of the PR and between f49e54b and c830bbc.

📒 Files selected for processing (10)
  • tests/constants.py
  • tests/e2e_shards.py
  • tests/test_end2end.py
  • tests/test_unittests.py
  • tools/conversion_registry.py
  • tools/modules/exporter.py
  • tools/version_detection/__init__.py
  • tools/version_detection/version_detection.py
  • tools/yolo/ultralytics
  • tools/yolo/yolo26_exporter.py

Included review availability: Your plan provides up to 1 included review per hour; 0 remain after this review.


📝 Walkthrough

Walkthrough

YOLO26 depth models now have dedicated version detection, registry wiring, exporter handling, map-output NNArchive generation, model fixtures, end-to-end coverage, and updated shard manifests.

Changes

YOLO26 depth conversion

Layer / File(s) Summary
Depth conversion detection and registration
tools/version_detection/..., tools/conversion_registry.py
Adds YOLOV26_DEPTH_CONVERSION, detects models containing yolo26 and DepthModel, exports the identifier, and registers it with the YOLO26 exporter.
Depth model export path
tools/yolo/ultralytics, tools/yolo/yolo26_exporter.py, tools/modules/exporter.py
Recognizes Ultralytics Depth heads, maps them to output_yolo26, and builds map-output NNArchives without min/max scaling.
Depth model test coverage
tests/constants.py, tests/test_end2end.py, tests/e2e_shards.py, tests/test_unittests.py
Adds the depth model fixture and NNArchive assertions, assigns the tests to E2E shards, and updates expected shard counts.

Estimated code review effort: 3 (Moderate) | ~20 minutes

Merge Risk: ⚪ Minimal · up to c830b

This adds YOLO26-Depth conversion and depth-map NNArchive output with matching end-to-end contract coverage. No actionable merge-readiness risk remains.

Sequence Diagram(s)

sequenceDiagram
  participant CLI
  participant VersionDetection
  participant YOLO26Exporter
  participant ArchiveGenerator
  CLI->>VersionDetection: Detect YOLO26 DepthModel
  VersionDetection-->>CLI: Return YOLOV26_DEPTH_CONVERSION
  CLI->>YOLO26Exporter: Load and export depth model
  YOLO26Exporter->>ArchiveGenerator: Build MapOutputParser archive
  ArchiveGenerator-->>CLI: Return output_yolo26 NNArchive
Loading

Suggested reviewers: dtronmans

🚥 Pre-merge checks | ✅ 4 | ❌ 1

❌ Failed checks (1 warning)

Check name Status Explanation Resolution
Docstring Coverage ⚠️ Warning Docstring coverage is 50.00% which is insufficient. The required threshold is 80.00%. Docstring coverage is scoped to functions touched by this diff. Analyzed 12 functions across 9 files. (1 skipped: … Write docstrings for the functions missing them to satisfy the coverage threshold.
✅ Passed checks (4 passed)
Check name Status Explanation
Description Check ✅ Passed Check skipped - CodeRabbit’s high-level summary is enabled.
Title check ✅ Passed The title clearly and concisely describes the main change: adding YOLO26 depth export support.
Linked Issues check ✅ Passed Check skipped because no linked issues were found for this pull request.
Out of Scope Changes check ✅ Passed Check skipped because no linked issues were found for this pull request.
Full details: Docstring Coverage

Explanation

Docstring coverage is 50.00% which is insufficient. The required threshold is 80.00%. Docstring coverage is scoped to functions touched by this diff. Analyzed 12 functions across 9 files. (1 skipped: 1 unsupported.)

✨ Finishing Touches 💡 1
📝 Generate docstrings 💡
  • Create stacked PR
  • Commit on current branch
🧪 Generate unit tests (beta)
  • Create PR with unit tests
  • Commit unit tests in branch feat/yolov26-depth-map-output

Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.

❤️ Share

Comment @coderabbitai help to get the list of available commands.

@github-actions

github-actions Bot commented Sep 7, 2026

Copy link
Copy Markdown

☂️ Python Coverage

current status: ✅

Overall Coverage

Lines Covered Coverage Threshold Status
2054 1135 55% 0% 🟢

New Files

No new covered files...

Modified Files

No covered modified files...

updated for commit: 39802c2 by action🐍

@github-actions

github-actions Bot commented Sep 7, 2026

Copy link
Copy Markdown

Test Results

  6 files    6 suites   14m 18s ⏱️
 34 tests  34 ✅ 0 💤 0 ❌
204 runs  204 ✅ 0 💤 0 ❌

Results for commit 39802c2.

@klemen1999 klemen1999 left a comment

Copy link
Copy Markdown
Collaborator

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Generally LGTM, just one small comment

Comment thread tools/yolo/yolo26_exporter.py
@rolandocortez
rolandocortez merged commit 30960ed into main Sep 8, 2026
26 checks passed
@rolandocortez
rolandocortez deleted the feat/yolov26-depth-map-output branch September 8, 2026 11:31
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

3 participants