I have a question about using YOLOv5 models with TensorFlow Lite.

[chung2324]

I'm new to machine learning and have been learning about object detection models. I trained a YOLOv5 model using Ultralytics and would like to deploy it on an edge device that only supports TensorFlow Lite models.

When I tried converting the trained YOLOv5 model to TFLite format, I am running into issues getting the converted model to work correctly on the edge device.

Does anyone have any tips or advice for a machine learning newbie like myself on properly converting a YOLOv5 model to TFLite for edge deployment?

Model:
Inputs float32 Shape[1,256,256,3]
Output float32 Shape [1,84,1360]