# SPDX-License-Identifier: Apache-2.0 # SPDX-FileCopyrightText: Copyright contributors to the vLLM project import importlib.machinery import os import sys import types # In symlink mode (VLLM_FLASH_ATTN_SRC_DIR), cute/ is a symlink to the real # source tree and its files use `flash_attn.cute.*` imports (not rewritten). # Register a virtual `flash_attn` package so those imports resolve. _cute_dir = os.path.join(os.path.dirname(__file__), "cute") if os.path.islink(_cute_dir) and "flash_attn" not in sys.modules: _fa_mod = types.ModuleType("flash_attn") _fa_mod.__path__ = [os.path.dirname(os.path.realpath(_cute_dir))] _fa_mod.__package__ = "flash_attn" _fa_mod.__spec__ = importlib.machinery.ModuleSpec( "flash_attn", None, is_package=True ) _fa_mod.__spec__.submodule_search_locations = _fa_mod.__path__ sys.modules["flash_attn"] = _fa_mod from vllm.vllm_flash_attn.flash_attn_interface import ( # noqa: E402 FA2_AVAILABLE, FA3_AVAILABLE, compile_flash_attn_varlen_func_from_specs, fa_version_unsupported_reason, flash_attn_varlen_func, get_scheduler_metadata, is_fa_version_supported, ) if not (FA2_AVAILABLE or FA3_AVAILABLE): raise ImportError( "vllm.vllm_flash_attn requires the CUDA flash attention extensions " "(_vllm_fa2_C or _vllm_fa3_C). On ROCm, use upstream flash_attn." ) __all__ = [ "compile_flash_attn_varlen_func_from_specs", "fa_version_unsupported_reason", "flash_attn_varlen_func", "get_scheduler_metadata", "is_fa_version_supported", ]