apps/FastDeploy
1
0
Fork 0
mirror of https://github.com/PaddlePaddle/FastDeploy.git synced 2026-05-06 15:40:33 +08:00
Code Issues Actions 2 Packages Projects Releases Wiki Activity

[Feature] Support redundant expert for eplb (#5918)

Browse Source 
* [BugFix] support redundant expert for eplb

* support redundant expert for eplb

* support redundant expert for eplb

* update

* fix ci eplb
This commit is contained in:
xiaoxiaohehe001
2026-01-09 17:13:24 +08:00
committed by GitHub
parent e6cdea4492
commit 00a01ae024
11 changed files with 36 additions and 17 deletions
Download Patch File Download Diff File Expand all files Collapse all files
fastdeploy/model_executor/layers/backends/xpu/moe/ep.py
+1 -1
View File
@@ -277,7 +277,7 @@ class XPUEPRunner:
moe_topk=self.top_k,
apply_norm_weight=True, # apply_norm_weight
enable_softmax_top_k_fused=False,
redundant_ep_rank_num_plus_one=layer.fd_config.model_config.redundant_experts_num + 1,
redundant_ep_rank_num_plus_one=layer.fd_config.eplb_config.redundant_experts_num + 1,
)
else:
topk_idx, topk_weights = fastdeploy.model_executor.ops.xpu.moe_topk_select(