Module quantization 

Quantization and Compression Functions

This module provides quantization operations for model compression including:

• Uniform and non-uniform quantization
• Dynamic quantization schemes
• Pruning utilities (magnitude-based, structured, unstructured)
• Model compression techniques
• Low-precision computation functions
• Knowledge distillation utilities

Enums

QuantizationScheme

Quantization schemes

QuantizationType

Quantization data types

Functions

dynamic_quantize

Dynamic quantization with automatic scale/zero-point calculation

fake_quantize

Quantize-aware training (QAT) simulation

gradual_magnitude_prune

Gradual magnitude pruning with sparsity scheduling

lottery_ticket_prune

Lottery ticket hypothesis: find winning subnetworks

magnitude_prune

Magnitude-based pruning

quantization_error_analysis

Quantization error analysis

uniform_dequantize

Dequantize uniformly quantized tensor

uniform_quantize

Uniform quantization

weight_clustering

Weight clustering for compression