Skip to content

Refactor YOLOv8+ export/quantize script in download_public_models +fix multistream sample model-proc + add YOLO11s, YOLO26s support in multistream sample#608

Merged
dmichalo merged 2 commits intomainfrom
dmichalo_PublicModelsSwitchToUltralytics
Feb 11, 2026
Merged

Refactor YOLOv8+ export/quantize script in download_public_models +fix multistream sample model-proc + add YOLO11s, YOLO26s support in multistream sample#608
dmichalo merged 2 commits intomainfrom
dmichalo_PublicModelsSwitchToUltralytics

Conversation

@dmichalo
Copy link
Contributor

@dmichalo dmichalo commented Feb 10, 2026

Description

Please include a summary of the changes and the related issue. List any dependencies that are required for this change.

Execution: https://github.com/open-edge-platform/dlstreamer/actions/runs/21877320003

Fixes # (issue)

Any Newly Introduced Dependencies

Please describe any newly introduced 3rd party dependencies in this change. List their name, license information and how they are used in the project.

How Has This Been Tested?

Please describe the tests that you ran to verify your changes. Provide instructions so we can reproduce. Please also list any relevant details for your test configuration

Checklist:

  • I agree to use the MIT license for my code changes.
  • I have not introduced any 3rd party components incompatible with MIT.
  • I have not included any company confidential information, trade secret, password or security token.
  • I have performed a self-review of my code.

Replace the large associative YOLO_MODELS map with a simple YOLO_V8_AND_LATER_MODELS array and rename the exporter to export_and_quantize_yolov8_and_later_model. Simplify the embedded Python export flow to use ultralytics' dynamic export for FP32/FP16 and INT8 (using model.export with dynamic/half/int8 flags) and move converted outputs into FP32/FP16/INT8 folders via os.rename. Remove manual OpenVINO Core read/modify/save steps and model_type-specific runtime-info/output renaming. Update the loop to iterate the new array and pass the quantize dataset parameter through; INT8 export still uses the provided YAML dataset.
…mple

Add new supported models (yolo11s, yolo26s) to the multi-stream sample and README. Generalize README note to instruct downloading required models to $MODELS_PATH/public/{model_name}/FP16 and list supported models plus an additional example. Update multi_stream_sample.sh help text and defaults to include the new models, add empty model-proc entries for them, and simplify/robustify model-proc handling: only call realpath and append model-proc when a model-proc file is set, otherwise omit the model-proc parameter from model arguments.
@dmichalo dmichalo force-pushed the dmichalo_PublicModelsSwitchToUltralytics branch from 25f7de7 to 7c19428 Compare February 10, 2026 15:10
@dmichalo dmichalo merged commit afa16eb into main Feb 11, 2026
199 of 259 checks passed
@dmichalo dmichalo deleted the dmichalo_PublicModelsSwitchToUltralytics branch February 11, 2026 17:29
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.

4 participants