Replace openvino-dev with OpenVINO Runtime inference

[GabrielDornelles]

One of OpenVINO use cases is to run it on Raspberry devices. I'm proposing to make the inference using openvino.runtime, which is installable at raspberry and is available at latest openvino package (tested with openvino==2022.1.0).

The previous inference using openvino-dev was not usable at devices where the dev package is not available.

🛠️ PR Summary

_{Made with ❤️ by Ultralytics Actions}

🌟 Summary

Improved OpenVINO integration in YOLOv5's model loading and inference.

📊 Key Changes

• Updated OpenVINO package requirement from openvino-dev to openvino.
• Switched to newer OpenVINO Python API usage with Core instead of IECore.
• Simplified OpenVINO network reading and compiling processes.
• Streamlined OpenVINO inference code to reduce verbosity and complexity.

🎯 Purpose & Impact

• Purpose: The update ensures compatibility with the latest OpenVINO version and takes advantage of the more user-friendly API.
• Impact: Users leveraging OpenVINO for model inference will experience a more straightforward setup with potentially enhanced stability and performance improvements. 🚀

[github-actions]

👋 Hello @GabrielDornelles, thank you for submitting a YOLOv5 🚀 PR! To allow your work to be integrated as seamlessly as possible, we advise you to:

• ✅ Verify your PR is up-to-date with upstream/master. If your PR is behind upstream/master an automatic GitHub Actions merge may be attempted by writing /rebase in a new comment, or by running the following code, replacing 'feature' with the name of your local branch:

git remote add upstream https://github.com/ultralytics/yolov5.git
git fetch upstream
# git checkout feature  # <--- replace 'feature' with local branch name
git merge upstream/master
git push -u origin -f

• ✅ Verify all Continuous Integration (CI) checks are passing.
• ✅ Reduce changes to the absolute minimum required 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

[glenn-jocher]

@GabrielDornelles thanks for the PR! Is there a way to use the openvino package for export also instead of openvino-dev?

yolov5/export.py

Lines 171 to 187 in 2b64b45

	def export_openvino(model, im, file, half, prefix=colorstr('OpenVINO:')):
	# YOLOv5 OpenVINO export
	try:
	check_requirements(('openvino-dev',)) # requires openvino-dev: https://pypi.org/project/openvino-dev/
	import openvino.inference_engine as ie
	LOGGER.info(f'\n{prefix} starting export with openvino {ie.__version__}...')
	f = str(file).replace('.pt', f'_openvino_model{os.sep}')
	cmd = f"mo --input_model {file.with_suffix('.onnx')} --output_dir {f} --data_type {'FP16' if half else 'FP32'}"
	subprocess.check_output(cmd, shell=True)
	LOGGER.info(f'{prefix} export success, saved as {f} ({file_size(f):.1f} MB)')
	return f
	except Exception as e:
	LOGGER.info(f'\n{prefix} export failure: {e}')

[GabrielDornelles]

@GabrielDornelles thanks for the PR! Is there a way to use the openvino package for export also instead of openvino-dev?

yolov5/export.py

Lines 171 to 187 in 2b64b45

	def export_openvino(model, im, file, half, prefix=colorstr('OpenVINO:')):
	# YOLOv5 OpenVINO export
	try:
	check_requirements(('openvino-dev',)) # requires openvino-dev: https://pypi.org/project/openvino-dev/
	import openvino.inference_engine as ie
	LOGGER.info(f'\n{prefix} starting export with openvino {ie.__version__}...')
	f = str(file).replace('.pt', f'_openvino_model{os.sep}')
	cmd = f"mo --input_model {file.with_suffix('.onnx')} --output_dir {f} --data_type {'FP16' if half else 'FP32'}"
	subprocess.check_output(cmd, shell=True)
	LOGGER.info(f'{prefix} export success, saved as {f} ({file_size(f):.1f} MB)')
	return f
	except Exception as e:
	LOGGER.info(f'\n{prefix} export failure: {e}')

I believe yes, I uninstalled dev package and the command works the same way in the non-dev package. The only thing is that it doesn't have a __version__ method, but it has a get_version that will display the following:

[glenn-jocher]

@GabrielDornelles looking at these two it seems that we still need openvino-dev for export:
https://pypi.org/project/openvino/
https://pypi.org/project/openvino-dev/

I'll take a closer look

[glenn-jocher]

@GabrielDornelles PR is merged. I validated PR and master exported OpenVINO models and saw same mAP. Also openvino-dev install provides for import openvino, so no changes needed by default for existing users.

Thank you for your contributions to YOLOv5 🚀 and Vision AI ⭐