Quant

UP supports the whole pipline of training and interfering;

Codes

Configs

It contains the illustration of common configs.

Repos

Dataset related modules

1. Dataset types:

• coco

2. The type of datasets can be chosen by setting ‘type’ in Dataset (default is coco). The config is as followed.

dataset:
  type: coco
  kwargs:
    ...

Quant setting

quant:
  ptq_only: False
  deploy_backend: tensorrt
  cali_batch_size: 900
  prepare_args:
    extra_qconfig_dict:
      w_observer: MinMaxObserver
      a_observer: EMAMinMaxObserver
      w_fakequantize: FixedFakeQuantize
      a_fakequantize: FixedFakeQuantize
    leaf_module: [Space2Depth, FrozenBatchNorm2d]
    extra_quantizer_dict:
      additional_module_type: [ConvFreezebn2d, ConvFreezebnReLU2d]