I have an object detection model based on YOLOV4 on my own dataset. I successfully converted the model to TFlite since I need to deploy it on iOS and Android.
I’m trying to follow this instruction:
However, I don’t know what changes I should apply. I would appreciate it if you could please guide me.
@Mojgan_m Welcome to tensorflow forum !
Here’s a general outline of the steps involved, please try below options and let us know if any of the below works for you:
Use the TensorFlow Lite CocoaPods library for integration. Follow Apple’s documentation and tutorials for setup and usage.
Include the TensorFlow Lite library in your Gradle dependencies. Refer to Google’s documentation and examples for integration.
Use the Interpreter class to load the TFLite model from the file.
Resize and normalize images to match the model’s input requirements. Apply any necessary color space conversions (e.g., RGB to BGR).
Pass the preprocessed image data to the model’s Interpreter . Extract bounding boxes, class scores, and other relevant outputs.
*Apply confidence thresholds to filter detections. Map class indices to label names. Render bounding boxes and labels on the image or video.
Let us know if this helps!