vllm.model_executor.layers.quantization.utils ¶
Modules:
-
flashinfer_fp4_moe–Utility helpers for NVFP4 + FlashInfer fused-MoE path
-
flashinfer_mxint4_moe–Utility helpers for MxInt4 + FlashInfer fused-MoE path
-
flashinfer_utils– -
fp8_utils– -
humming_utils– -
int8_utils– -
machete_utils– -
marlin_utils– -
marlin_utils_fp4– -
marlin_utils_fp8– -
marlin_utils_test–Utility functions used for tests and benchmarks
-
mxfp4_utils– -
mxfp8_utils– -
nvfp4_emulation_utils– -
nvfp4_utils– -
quant_utils–This file is used for /tests and /benchmarks