apps/FastDeploy
1
0
Fork 0
mirror of https://github.com/PaddlePaddle/FastDeploy.git synced 2026-04-23 00:17:25 +08:00
Code Issues Actions 19 Packages Projects Releases Wiki Activity

[BugFix] Fix tbo nan (#6439)

Browse Source
This commit is contained in:
RichardWooSJTU
2026-03-02 14:28:48 +08:00
committed by GitHub
parent fe0b3a90ee
commit 7bd86f99a5
1 changed files with 0 additions and 4 deletions
Download Patch File Download Diff File Expand all files Collapse all files
fastdeploy/model_executor/layers/moe/fused_moe_deepgemm_backend.py
-4
View File
@@ -348,7 +348,6 @@ class DeepGemmFusedMoeMethod(MoEMethodBase):
ffn_out,
m_indices,
)
del permute_input
# swiglu
ffn_out = paddle.incubate.nn.functional.swiglu(ffn_out, None)
@@ -367,7 +366,6 @@ class DeepGemmFusedMoeMethod(MoEMethodBase):
)
ffn_in_x_scale_tensor = ffn_in_x_scale_tensor.T[: ffn_in_x.shape[0]]
del ffn_out
ffn_out = paddle.empty(
(token_all_num, getattr(layer, self.added_weight_attrs[1]).shape[1]),
dtype=paddle.bfloat16,
@@ -379,7 +377,6 @@ class DeepGemmFusedMoeMethod(MoEMethodBase):
ffn_out,
m_indices,
)
del ffn_in_x
# prmt back per rank
tmp_ffn_out = fastdeploy.model_executor.ops.gpu.ep_moe_expert_combine(
@@ -391,7 +388,6 @@ class DeepGemmFusedMoeMethod(MoEMethodBase):
False, # norm_topk_prob
1.0,
)
del ffn_out
else:
tmp_ffn_out = paddle.empty([0, hidden_size], paddle.bfloat16)