Use is_flashinfer_available to replace is_hip for flashinfer check (#1596)

Co-authored-by: Zhang Liangang <liangang.zhang@intel.com>
This commit is contained in:
Lianmin Zheng
2024-10-06 22:54:05 -07:00
committed by GitHub
parent 565b05f02f
commit 6a5b352aaf
9 changed files with 29 additions and 28 deletions

View File

@@ -47,10 +47,9 @@ from sglang.srt.layers.quantization.base_config import QuantizationConfig
from sglang.srt.layers.radix_attention import RadixAttention
from sglang.srt.managers.schedule_batch import global_server_args_dict
from sglang.srt.model_executor.forward_batch_info import ForwardBatch
from sglang.srt.utils import is_hip
from sglang.srt.utils import is_flashinfer_available
# ROCm: flashinfer available later
if not is_hip():
if is_flashinfer_available():
from flashinfer import bmm_fp8

View File

@@ -43,10 +43,9 @@ from sglang.srt.layers.quantization.base_config import QuantizationConfig
from sglang.srt.layers.radix_attention import RadixAttention
from sglang.srt.managers.schedule_batch import global_server_args_dict
from sglang.srt.model_executor.forward_batch_info import ForwardBatch
from sglang.srt.utils import is_hip
from sglang.srt.utils import is_flashinfer_available
# ROCm: flashinfer available later
if not is_hip():
if is_flashinfer_available():
from flashinfer import bmm_fp8