AIMET Installation

Quick Install

The AIMET PyTorch GPU PyPI packages are available for environments that meet the following requirements:

  • 64-bit Intel x86-compatible processor

  • Linux Ubuntu 22.04 LTS [Python 3.10] or Ubuntu 20.04 LTS [Python 3.8]

  • CUDA 12.0

  • Torch 2.2.2

Pip install

apt-get install liblapacke
python3 -m pip install aimet-torch

Release Packages

For other AIMET variants, install the latest version from the .whl files hosted at https://github.com/quic/aimet/releases

PyTorch

# Pytorch 2.1 with CUDA 12.x
python3 -m pip install https://github.com/quic/aimet/releases/download/1.34.0/aimet_torch-1.34.0.cu121-cp310-cp310-manylinux_2_34_x86_64.whl

# Pytorch 2.1 CPU only
python3 -m pip install https://github.com/quic/aimet/releases/download/1.34.0/aimet_torch-1.34.0.cpu-cp310-cp310-manylinux_2_34_x86_64.whl

# Pytorch 1.13 with CUDA 11.x
python3 -m pip install https://github.com/quic/aimet/releases/download/1.34.0/aimet_torch-1.34.0.cu117-cp310-cp310-manylinux_2_34_x86_64.whl

TensorFlow

# Tensorflow 2.10 GPU with CUDA 11.x
python3 -m pip install https://github.com/quic/aimet/releases/download/1.34.0/aimet_tensorflow-1.34.0.cu118-cp310-cp310-manylinux_2_34_x86_64.whl

# Tensorflow 2.10 CPU only
python3 -m pip install https://github.com/quic/aimet/releases/download/1.34.0/aimet_tensorflow-1.34.0.cpu-cp310-cp310-manylinux_2_34_x86_64.whl

Onnx

# ONNX 1.16 GPU with CUDA 11.x
python3 -m pip install https://github.com/quic/aimet/releases/download/1.34.0/aimet_onnx-1.34.0.cu117-cp310-cp310-manylinux_2_34_x86_64.whl

# ONNX 1.16 CPU
python3 -m pip install https://github.com/quic/aimet/releases/download/1.34.0/aimet_onnx-1.34.0.cpu-cp310-cp310-manylinux_2_34_x86_64.whl

For older versions, please browse the releases at https://github.com/quic/aimet/releases and follow the documentation corresponding to that release to select and install the appropriate package.

System Requirements

The AIMET package requires the following host platform setup:

  • 64-bit Intel x86-compatible processor

  • Linux Ubuntu: 22.04 LTS

  • bash command shell

  • For GPU variants:

To use the GPU accelerated training modules an Nvidia CUDA enabled GPU with a minimum Nvidia driver version of 455+ is required. Using the latest driver is always recommended, especially if using a newer GPU. Both CUDA and cuDNN (the more advanced CUDA interface) enabled GPUs are supported.

Advanced Installation Instructions

There are two ways to setup and install AIMET:

  • On your host machine

  • Using our pre-built development Docker images

Please click on the appropriate link for installation instructions: