ModelSDK APIs

The Modal SDK supports Post Training Quantization, Quantization Aware Training and Model Graph Surgery through the following APIs.

Post Training Quantization (PTQ)

Available through the afe.apis.loaded_net, afe.apis.defines and afe.apis.model module from the afe package.

Quantization Aware Training (QAT)

Available through the sima_qat.qat_api module.

Graph Surgery APIs

Available through the onnx_helpers package.