apps/FastDeploy
1
0
Fork 0
mirror of https://github.com/PaddlePaddle/FastDeploy.git synced 2025-10-29 02:52:55 +08:00
Code Issues Actions 5 Packages Projects Releases Wiki Activity

[MTP]fix speculate-decoding in dpep mode (#4351)
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
This commit is contained in:
freeliuzc
2025-10-11 17:16:57 +08:00
committed by GitHub
parent 0c4c28d799
commit 8d629568f2
1 changed files with 7 additions and 1 deletions
Download Patch File Download Diff File Expand all files Collapse all files

8
fastdeploy/output/token_processor.py
View File

@@ -156,7 +156,13 @@ class TokenProcessor:
try: try:
is_blocking = True is_blocking = True
if self.speculative_decoding: if self.speculative_decoding:
speculate_get_output(self.output_tokens, rank_id, is_blocking, False) if (
self.cfg.parallel_config.enable_expert_parallel
and self.cfg.parallel_config.data_parallel_size > 1
):
speculate_get_output(self.output_tokens, rank_id, is_blocking, True)
else:
speculate_get_output(self.output_tokens, rank_id, is_blocking, False)
if self.output_tokens[0] == -2: if self.output_tokens[0] == -2:
continue continue