(feat) Parallel loading and execution models WIP

[pinball83]

🛠️ PR Summary

_{Made with ❤️ by Ultralytics Actions}

WARNING ⚠️ this PR is very large, summary may not cover all changes.

🌟 Summary

Multi-model refactor enabling concurrent model loading/inference across Android and iOS, with unified overlays, improved streaming payloads, and example app updates to use multiple YOLO models. 🎛️📱

📊 Key Changes

• Android: Added multi-model APIs (setModels) in YOLOPlatformView/YOLOView with parallel inference, shared executor, model generation guards, and unified overlay rendering for DETECT/SEGMENT/POSE/OBB.
• Android: Enhanced streaming payloads with per-detection modelName, viewBoundingBox, combined results aggregation, throttling/frequency controls, and resilience for event channel reconnection.
• Android: ObjectDetector small refactors (formatting, logs, method signature readability) and safer label loading logs.
• iOS: SwiftYOLOPlatformView supports models args and adds setModels (initial support + first-model fallback), retains thresholds/UI controls syncing, and streaming setup.
• iOS: VideoCapture now more robust in selecting camera devices across hardware types; safer setup and cleanup.
• Example app: Controller/UI refactor to support multiple active models via YOLOModelSpec, dynamic toggling, and updated YOLOView invocation; simplified overlays to use controller-provided models.
• iOS project: Updated Podfile.lock checksums and Xcode project Development Team ID.

🎯 Purpose & Impact

• Enables concurrent tasks (e.g., segment + pose) in one camera pipeline, improving flexibility and demo quality.
• Delivers richer, more UI-ready streaming data (normalized, pixel, and view-mapped boxes plus model attribution).
• Improves stability and performance with parallel inference, output throttling, and reconnection retries.
• Simplifies app integration via a clearer model spec (YOLOModelSpec) and unified overlay rendering for multi-model results.
• Broadens device compatibility and reliability on iOS with improved camera device selection.

[github-actions]

All Contributors have signed the CLA. ✅
_{Posted by the CLA Assistant Lite bot.}

[UltralyticsAssistant]

👋 Hello @pinball83, thank you for submitting a ultralytics/yolo-flutter-app 🚀 PR! To ensure a seamless integration of your work, please review the following checklist:

• ✅ Define a Purpose: Clearly explain the purpose of your fix or feature in your PR description, and link to any relevant issues. Ensure your commit messages are clear, concise, and adhere to the project's conventions.
• ✅ Synchronize with Source: Confirm your PR is synchronized with the ultralytics/yolo-flutter-app main branch. If it's behind, update it by clicking the 'Update branch' button or by running git pull and git merge main locally.
• ✅ Ensure CI Checks Pass: Verify all Ultralytics Continuous Integration (CI) checks are passing. If any checks fail, please address the issues.
• ✅ Update Documentation: Update the relevant documentation for any new or modified features.
• ✅ Add Tests: If applicable, include or update tests to cover your changes, and confirm that all tests are passing.
• ✅ Sign the CLA: Please ensure you have signed our Contributor License Agreement if this is your first Ultralytics PR by writing "I have read the CLA Document and I sign the CLA" in a new message.
• ✅ Minimize Changes: Limit your changes to the minimum necessary for your bug fix or feature addition. "It is not daily increase but daily decrease, hack away the unessential. The closer to the source, the less wastage there is." — Bruce Lee

For more guidance, please refer to our Contributing Guide. This is an automated message and an engineer will be with you shortly. Thanks for contributing to Ultralytics! ✨🚀

[UltralyticsAssistant]

🔍 PR Review

_{Made with ❤️ by Ultralytics Actions}

The Android platform view no longer loads a model when only the legacy modelPath/task parameters are passed, which will break existing apps; we should restore the single-model fallback in the initializer.

💬 Posted 1 inline comment

⚠️ Large PR: Review focused on critical issues. Some details may not be covered.

📋 Skipped 1 file (lock files, minified, images, etc.)

[UltralyticsAssistant]

⚠️ HIGH: creationParams used to rely on modelPath/task for single-model loading. With this branch we now skip setModel entirely whenever models isn’t supplied, so every existing caller that still sends the legacy params will end up with a camera preview and no inference. We need to keep the old fallback (e.g. call setModel(...) here) so the API remains backwards compatible.

[pinball83]

I have read the CLA Document and I sign the CLA

[codecov]

Codecov Report

❌ Patch coverage is 45.08197% with 67 lines in your changes missing coverage. Please review.

Files with missing lines	Patch %	Lines
lib/yolo_view.dart	1.78%	55 Missing ⚠️
lib/widgets/yolo_controller.dart	0.00%	5 Missing ⚠️
lib/models/yolo_model_spec.dart	91.48%	4 Missing ⚠️
lib/models/yolo_result.dart	72.72%	3 Missing ⚠️

📢 Thoughts on this report? Let us know!

[asabri97]

@pinball83 is this PR ready for review?

[pinball83]

@pinball83 is this PR ready for review?

not yet actually, I need to test ios part application so I've been wait chance to get iphone, so now I have one after couple day I finalise PR

[pderrenger]

Thanks for the update and for signing the CLA—please mark this PR as Draft while you finish iOS testing; when ready, add a brief device/OS and FPS/latency matrix to the PR description, consider splitting platform changes from example‑app updates to keep review focused, then tag me to begin review; see the Development Workflow for expectations. See the Development Workflow for expectations.