Sample List

**Table 1** Samples
Framework	Feature	How to Obtain
PyTorch	Accuracy-based Automatic Quantization	Click auto_calibration to obtain the sample. For details, see the readme file.
	Uniform Quantization	Click calibration to obtain the sample. For details, see the readme file.
	NUQ	Click calibration_nuq to obtain the sample. For details, see the readme file.
	QAT	Click retrain to obtain the sample. For details, see the readme file.
	Auto Channel Pruning Search	Click auto_channel_prune to obtain the sample. For details, see the readme file.
	Filter-Level Sparsity	Click channel_prune to obtain the sample. For details, see the readme file.
	2:4 Structured Sparsity	Click selective_prune to obtain the sample. For details, see the readme file.
	Compression Combination	Click mix_compression to obtain the sample. For details, see the readme file.
	Tensor Decomposition	Click tensor_decompose to obtain the sample. For details, see the readme file.
	QAT in Single-Operator Mode	Click retrain_qat_op to obtain the sample. For details, see the readme file.
	Layer-wise Distillation	Click distillation to obtain the sample. For details, see the readme file.
	ADA Adaptive Rounding Quantization	Click ada_round_calibration to obtain the sample. For details, see the readme file.
PyTorch	KV Cache Quantization	Click kv_cache_quantization to obtain the sample. For details, see the readme file.
ONNX	CLI-based Quantization PTQ using the CLI QAT model adaptation to CANN format using the CLI	Click cmd to obtain the sample. For details, see the readme file.
	Accuracy-based Automatic Quantization	Click accuracy_based_auto_calibration to obtain the sample. For details, see the readme file for accuracy-based automatic quantization.
	Uniform Quantization	Click calibration to obtain the sample. For details, see the readme file for uniform quantization.
	NUQ	Click calibration_nuq to obtain the sample. For details, see the readme file for non-uniform quantization.
	QAT Model Adaptation to CANN Format	Click convert_qat2ascend to obtain the sample. For details, see the readme file for converting a QAT model to a CANN model.
TensorFlow	CLI-based Quantization PTQ using the CLI QAT model adaptation to CANN format using the CLI	Click cmd to obtain the sample. For details, see the readme file.
	Accuracy-based Automatic Quantization	Click auto_calibration to obtain the sample. For details, see the readme file.
	Uniform Quantization	Click calibration to obtain the sample. For details, see the readme file.
	NUQ	Click calibration_nuq to obtain the sample. For details, see the readme file.
	QAT	Click retrain to obtain the sample. For details, see the readme file.
	Auto Channel Pruning Search	Click auto_channel_prune to obtain the sample. For details, see the readme file.
	Filter-Level Sparsity (Manual Sparsity)	Click channel_prune to obtain the sample. For details, see the readme file.
	2:4 Structured Sparsity	Click selective_prune to obtain the sample. For details, see the readme file.
	Compression Combination	Click mix_compression to obtain the sample. For details, see the readme file.
	Tensor Decomposition	Click tensor_decompose to obtain the sample. For details, see the readme file.
	Model Adaptation Using convert_model API	Click convert_model to obtain the sample. For details, see the readme file.
	QAT Model Adaptation to CANN Format	Click convert_qat2ascend to obtain the sample. For details, see the readme file.
Caffe	CLI-based Quantization	Click cmd to obtain the sample. For details, see the readme file.
	Accuracy-based Automatic Quantization	Click auto_calibration to obtain the sample. For details, see the readme file.
	Uniform Quantization	Click calibration to obtain the sample. For details, see the readme file.
	NUQ	Automatic non-uniform quantization: Click auto_calibration_nuq to obtain the sample. For details, see the readme file. Static non-uniform quantization: Click calibration_nuq to obtain the sample. For details, see the readme file.
	QAT	Click retrain to obtain the sample. For details, see the readme file.
	Tensor Decomposition	Click tensor_decompose to obtain the sample. For details, see the readme file.
	Model Adaptation	Click convert_model to obtain the sample. For details, see the readme file.
TensorFlow, Ascend	MobileNetV2	Post-training quantization of the classification network model. Click amct_tensorflow_ascend to obtain the sample from the mobilenetv2 directory. For details, see the readme file.
TensorFlow, Ascend	YOLOv3	Post-training quantization of the detection network model. Click amct_tensorflow_ascend to obtain the sample from the yolov3 directory. For details, see the readme file.