Horizon OpenExplorer currently provides 2 sets of model quantization schemes at the same time.

• PTQ: Post-training Quantization.
• QAT: Quantization Awareness Training, which only supports the Pytorch framework for now.

Both solutions do not interfere with the training phase of the floating-point model, which is your own responsibility. Horizon has also open-sourced some reference PyTorch implementations for tasks such as classification, detection, and segmentation under the samples/ directory of the OE package, which can be used for reference and support training and reproduction on the host machine.

For both of the above quantization schemes, Horizon provides both local manual installation and Docker containers. We strongly recommend using Docker containers as they do not pollute the local environment and is easy to use.

We will introduce you to the development and running environment requirements and deployment process required for using the toolchain in two parts: Pre-installation Preparation and Software Installation.