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

[Metax] optimize flash mla (#4915)

Browse Source
This commit is contained in:
xiaozude
2025-11-12 16:43:46 +08:00
committed by GitHub
parent 9d9f5df8d0
commit c45b3ccb52
5 changed files with 37 additions and 38 deletions
Download Patch File Download Diff File Expand all files Collapse all files
fastdeploy/model_executor/models/deepseek_v3.py
+1 -1
View File
@@ -728,7 +728,7 @@ class DeepseekV3ForCausalLM(ModelForCasualLM):
seq_lens_decoder = forward_meta.seq_lens_decoder
seq_lens_this_time = forward_meta.seq_lens_this_time
current_total_tokens = paddle.sum(seq_lens_this_time)
current_total_tokens = forward_meta.ids_remove_padding.shape[0]
position_ids = self.position_ids_buffer[:current_total_tokens]
mask_encoder_batch = self.mask_encoder_batch_buffer[:current_total_tokens]