[Doc] Consistent naming of attention backends (#9498)

Signed-off-by: Thomas Parnell <tpa@zurich.ibm.com>
This commit is contained in:
Thomas Parnell
2024-10-21 22:29:57 +08:00
committed by GitHub
parent 696b01af8f
commit 496e991da8
14 changed files with 23 additions and 19 deletions
+2 -2
View File
@@ -29,8 +29,8 @@ if TYPE_CHECKING:
logger = init_logger(__name__)
MULTI_STEP_ATTENTION_BACKENDS = ["flash-attn", "rocm-flash-attn", "flashinfer"]
MULTI_STEP_CHUNKED_PREFILL_ATTENTION_BACKENDS = ["flash-attn"]
MULTI_STEP_ATTENTION_BACKENDS = ["FLASH_ATTN", "ROCM_FLASH", "FLASHINFER"]
MULTI_STEP_CHUNKED_PREFILL_ATTENTION_BACKENDS = ["FLASH_ATTN"]
def _get_supported_attention_backends(chunked_prefill_enabled: bool) \
-> List[str]: