backends

Modules

modelopt.torch.quantization.backends.fp8_per_tensor_gemm(...)	GEMM function for fp8 per tensor quantization.
modelopt.torch.quantization.backends.gemm_registry	Registry for specialized GEMM (General Matrix Multiplication) implementations.
modelopt.torch.quantization.backends.nvfp4_gemm(...)	GEMM function for fp4 quantization.
modelopt.torch.quantization.backends.utils	This file contains utility functions used by the quantization backend.

Quantization backends.