This version add some scripts to use your own dataset and a working config (in the configuration directory) you can build using docker or podman

podman build -t coral-tutorials https://github.com/google-coral/tutorials.git#:docker/object_detection

These Jupyter notebooks provide interactive Python tutorials for development with Coral. You can download these files and run them on a local Jupyter notebook, but they're designed for execution with Google Colab, using the links below.

Train and build a model for the Edge TPU:

• Retrain a classification model using post-training quant (TF2)

  This notebook builds the MobileNet V2 classification model using Keras (with TF2), retrains it to classify different flowers, quantizes it with post- training quantization, and then compiles it for the Edge TPU.

  We have another version of this tutorial using TF1.

• Retrain a classification model using quant-aware training (TF1)

  This notebook retrains a quant-aware version of MobileNet V1 (with TF1) to classify different flowers, and then compiles it for the Edge TPU.

• Retrain an object detection model (MobileNet) using quant-aware training (TF1)

  This notebook retrains a quant-aware version of MobileNet V1 SSD (with TF1) to detect cats and dogs, and then compiles it for the Edge TPU.

• Retrain an object detection model (MobileDet) using post-training quant (TF1)

  This notebook retrains the SSDLite MobileDet model to detect cats and dogs, quantizes it with post-training quantization, and then compiles it for the Edge TPU.

• Retrain an object detection model (EfficientDet) using TF Lite Model Maker (TF2)

  This notebook retrains the EfficientDet-Lite model using the TensorFlow Lite Model Maker library (based on TF2) to detect different salad ingredients, exports it to TF Lite (using post-training quantization), and then compiles it for the Edge TPU.

• Build and train an LSTM time series model (TF2)

  This notebook shows how to create an LSTM model that makes weather predictions, and then quantize the model and compile it for the Edge TPU.

Other tutorials:

• Web-based Edge TPU Compiler

  This Colab compiles a TensorFlow Lite model for the Edge TPU, in case you don't have a system that's compatible with the Edge TPU Compiler (Debian Linux only).

• Run Colab on a Coral Dev Board

  This shows how to run a Jupyter notebook on your Dev Board from a Google Colab interface on your host computer.

• Build the libcoral C++ examples

  This is a convenient way to build all the C++ examples in github.com/google-coral/libcoral, and download them to your computer.

• Build the C++ "lstpu" example

  This is a convenient way to build the "lstpu" C++ example (it lists all Edge TPUs in your system), and download it to your computer.