Google Colaboratory (Colab) Notebooks

These Colab notebooks contain interactive sample applications using IREE's Python bindings and ML framework integrations.

Notebooks

Framework	Notebook file	Description	Link
Generic	low_level_invoke_function	Shows low level IREE python binding concepts
PyTorch	pytorch_aot_advanced	Uses Turbine to export a PyTorch program
PyTorch	pytorch_aot_simple	Uses Turbine to export a simple PyTorch program
PyTorch	pytorch_jit	Uses Turbine for eager execution in a PyTorch session
PyTorch	pytorch_huggingface_whisper	Imports the Whisper model from Hugging Face
TensorFlow	tensorflow_edge_detection	Performs image edge detection
TensorFlow	tensorflow_hub_import	Imports a MobileNet V2 model from TensorFlow Hub
TensorFlow	tensorflow_mnist_training	Compiles, trains, and executes a neural network
TensorFlow	tensorflow_resnet	Runs a pretrained ResNet50 model
TFLite	tflite_text_classification	Runs a pretrained text classification model

Working with GitHub

Refer to Colab's GitHub demo for general information about using Colab with GitHub.

To make changes to a notebook in this repository, one possible workflow is:

• Open or create the notebook in Colab
• Connect to a hosted or local runtime
• Make your changes, run the notebook, etc.
• Download the modified notebook using File > Download .ipynb
• Move the downloaded notebook file into a clone of this repository and submit a pull request

Testing

This notebooks are tested continuously by the samples.yml CI job.