Welcome to AI Model Efficiency Toolkit API Docs!

AI Model Efficiency Toolkit (AIMET) is a software toolkit that enables users to compress and quantize ML models. The resulting models returned by AIMET can be further trained (or fine-tuned) to dramatically improve accuracy lost due to quantization and compression.

AIMET is designed to work generically on any user-provided model. At present, AIMET supports TensorFlow, Keras, PyTorch, and ONNX frameworks.

Please follow the links below to see AIMET APIs for either PyTorch, TensorFlow, Keras, or ONNX.

Note

AIMET Keras API requires Tensorflow 2.4 or later.

• AIMET APIs for PyTorch
  • PyTorch Model Quantization API
  • PyTorch Model Compression API
  • PyTorch Model Visualization API for Compression
  • PyTorch Model Visualization API for Quantization
  • PyTorch Debug API

• AIMET APIs for TensorFlow
  • TensorFlow Model Guidelines
  • TensorFlow Model Quantization API
  • TensorFlow Model Compression API
  • TensorFlow Model Visualization API for Quantization
  • Using AIMET Tensorflow APIs with Keras Models
  • Tensorflow Debug API

• AIMET APIs for Keras
  • Keras Model Quantization API
  • Keras Debug API
  • Keras Model Compression API

• AIMET APIs for ONNX
  • ONNX Model Quantization API
  • ONNX Debug API

Note

This documentation is auto-generated from the AIMET codebase using Sphinx

Indices and tables

• Index

• Module Index

• Search Page