modelopt.torch.quantization.backends.nvfp4_gemm

nvfp4_gemm(quant_module, input_tensor, bias=None)

GEMM function for fp4 quantization.