--model

Description

Sets the path (including the file name) of the model file to be quantized.

See Also

For Caffe network models, this option must be used together with --weights.

Argument

Argument: path (including the file name) of the model file.

Format: The path (including the file name) can contain letters, digits, underscores (_), hyphens (-), and periods (.).

Suggestions and Benefits

None

Example

• Caffe network model:

  --model=$HOME/test/ResNet-50-deploy.prototxt --weights=$HOME/test/ResNet-50-model.caffemodel

• TensorFlow network model:

  --model=$HOME/test/resnet_v1_50.pb

• ONNX network model:

  --model=$HOME/test/resnet101_v11.onnx

Restrictions

This option is required in both PTQ and QAT model adaptation to CANN format.