End-to-end Object Detection in TensorFlow Lite
This project shows how to train a custom detection model with the TFOD API, optimize it with TFLite, and perform inference with the optimized model.
object-detection tensorflow tensorflow-lite computer-vision code tutorial

The project shows how to train a custom detection model with the TFOD API (TF2 and TF1), optimize it with TFLite, and perform inference with the optimized model. It contains the following notebooks (pardon the naming) -

  • Training_a_pets_detector_model_within_minutes_with_TFOD_API.ipynb: Shows how to train a custom object detection model on the Pets dataset (non-eager mode) with Cloud TPUs. Note that it does not use TPUs offered by Colab.
  • Running_inference_with_a_custom_TFOD_API_model.ipynb: Shows how to export a SavedModel graph from the trained checkpoint files, and run inference.
  • Object_Detection_in_TFLite.ipynb: Shows how to quantize the original model, generate a TFLite model, and run inference.
  • Training_MobileDet_Custom_Dataset.ipynb: Shows how to train a custom object detection model on the Pets dataset (non-eager mode) on Colab (GPU), optimize the fine-tuned model with TFLite, and perform inference with the optimized model.

Don't forget to tag @sayakpaul in your comment, otherwise they may not be notified.

Authors original post
Calling `model.fit()` @ https://pyimagesearch.com | Netflix Nerd
Share this project
Similar projects
YOLOv4 With TensorFlow
YOLOv4, YOLOv4-tiny, YOLOv3, YOLOv3-tiny Implemented in Tensorflow 2.0, Android. Convert YOLO v4 .weights tensorflow, tensorrt and tflite.
Tensorflow2 Object Detection Tutorial
In this tutorial, we will be going step by step the complete training process of Tensorflow2 Object Detection.
Object Detection with RetinaNet
Implementing RetinaNet: Focal Loss for Dense Object Detection.
Real-time text detection with EAST in TFLite
Demonstrates the conversion process from the original EAST model to TFLite and how to use it on static images and also on real-time video feeds.