This Pull Request introduces an example of how to convert a YoloV5s model to half precision and use it on the Panorama device.
The change is inside the existing PT37_opengpu sample. This sample was only using an FP32 model but will now introduce the ability to convert this model to FP16 and then use it on the Panorama device. This model should take a dynamic input batch size.
We will be updating the inference latency metrics soon for this half precision model with various batch sizes
By submitting this pull request, I confirm that you can use, modify, copy, and redistribute this contribution, under the terms of your choice.
Issue #, if available:
Description of changes:
This Pull Request introduces an example of how to convert a YoloV5s model to half precision and use it on the Panorama device.
The change is inside the existing PT37_opengpu sample. This sample was only using an FP32 model but will now introduce the ability to convert this model to FP16 and then use it on the Panorama device. This model should take a dynamic input batch size.
We will be updating the inference latency metrics soon for this half precision model with various batch sizes
By submitting this pull request, I confirm that you can use, modify, copy, and redistribute this contribution, under the terms of your choice.