Skip to content

This repository shows how to train a custom detection model with the TFOD API, optimize it with TFLite, and perform inference with the optimized model.

License

Notifications You must be signed in to change notification settings

sayakpaul/E2E-Object-Detection-in-TFLite

Repository files navigation

E2E-Object-Detection-in-TFLite

This repository 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.

About the notebooks

Note

Training_a_pets_detector_model_within_minutes_with_TFOD_API.ipynb notebook uses Colab to demonstrate the training workflow but does not actually uses the Colab runtime for training. It uses Cloud TPUs.

Model usage on mobile devices

Sample applications (both Android and iOS) are available here.

TFLite model files

Available here (currently available in dynamic-range only). It was generated from the first notebook (Training_a_pets_detector_model_within_minutes_with_TFOD_API.ipynb).

Acknowledgement

I would like to thank the ML-GDE program for providing with GCP credits without which this project wouldn't have been possible.

About

This repository shows how to train a custom detection model with the TFOD API, optimize it with TFLite, and perform inference with the optimized model.

Topics

Resources

License

Stars

Watchers

Forks

Packages

No packages published