apps/FastDeploy
1
0
Fork 0
mirror of https://github.com/PaddlePaddle/FastDeploy.git synced 2025-10-28 02:42:21 +08:00
Code Issues Actions 5 Packages Projects Releases Wiki Activity

【BugFif】fix ep decode (#4138)
Some checks failed
CE Compile Job / ce_job_pre_check (push) Has been cancelled
Details
CE Compile Job / print_ce_job_pre_check_outputs (push) Has been cancelled
Details
CE Compile Job / FD-Clone-Linux (push) Has been cancelled
Details
CE Compile Job / Show Code Archive Output (push) Has been cancelled
Details
CE Compile Job / BUILD_SM8090 (push) Has been cancelled
Details
CE Compile Job / BUILD_SM8689 (push) Has been cancelled
Details
CE Compile Job / CE_UPLOAD (push) Has been cancelled
Details

Browse Source 
* suppoort expert num 3 pre rank

* fix ep decode
This commit is contained in:
gaoziyuan
2025-09-17 14:18:31 +08:00
committed by GitHub
parent 25aa2d94aa
commit 5027ed7239
1 changed files with 1 additions and 1 deletions
Download Patch File Download Diff File Expand all files Collapse all files

2
fastdeploy/model_executor/layers/moe/fused_moe_backend_base.py
View File

@@ -166,7 +166,7 @@ class MoEMethodBase(QuantMethodBase):
else:
if layer.fd_config.parallel_config.splitwise_role == "mixed":
self.ep_decoder_runner.clean_low_latency_buffer()
return self.apply_ep_prefill(layer, x, gate)
return self.apply_ep_decode(layer, x, gate)
else:
return self.apply_tp(layer, x, gate)