apps/FastDeploy
1
0
Fork 0
mirror of https://github.com/PaddlePaddle/FastDeploy.git synced 2025-10-05 16:48:03 +08:00
Code Issues Actions 2 Packages Projects Releases Wiki Activity

[feat] support fa3 backend for pd disaggregated (#2695)
Some checks failed
Deploy GitHub Pages / deploy (push) Has been cancelled
Details

Browse Source 
* support fa3 backend run in pd disaggregated

* support fa3 backend run in pd disaggregated

* support fa3 backend run in pd disaggregated

* support fa3 backend run in pd disaggregated

* delete use_fast_ffn
This commit is contained in:
Yuanle Liu
2025-07-03 22:33:27 +08:00
committed by GitHub
parent 00863c43fd
commit 240bdac2a4
26 changed files with 455 additions and 139 deletions
Download Patch File Download Diff File Expand all files Collapse all files

8
fastdeploy/platforms/cuda.py
View File

@@ -13,9 +13,6 @@
# See the License for the specific language governing permissions and
# limitations under the License.
"""
"""
cuda platform file
"""
import paddle
@@ -65,6 +62,11 @@ class CUDAPlatform(Platform):
return (
"fastdeploy.model_executor.layers.attention.MLAAttentionBackend"
)
elif selected_backend == _Backend.FLASH_ATTN:
logger.info("Using FLASH ATTN backend.")
return (
"fastdeploy.model_executor.layers.attention.FlashAttentionBackend"
)
else:
raise ValueError(
"Invalid attention backend you specified.\n"