apps/FastDeploy
1
0
Fork 0
mirror of https://github.com/PaddlePaddle/FastDeploy.git synced 2025-10-06 09:07:10 +08:00
Code Issues Actions 2 Packages Projects Releases Wiki Activity

Revert "[BugFix] fix ep (#3290)" (#3317)

Browse Source 
This reverts commit 86ff68be4b.
This commit is contained in:
Jiang-Jia-Jun
2025-08-11 16:17:58 +08:00
committed by GitHub
parent c000cff744
commit ca4e4ab911
2 changed files with 3 additions and 4 deletions
Download Patch File Download Diff File Expand all files Collapse all files

4
fastdeploy/worker/gpu_model_runner.py
View File

@@ -780,7 +780,7 @@ class GPUModelRunner(ModelRunnerBase):
output_padding_offset,
) = pre_process(
self.share_inputs["input_ids"],
getattr(self.share_inputs, "seq_lens_this_time", self.seq_lens_this_time_buffer),
self.share_inputs["seq_lens_this_time"],
self.speculative_decoding,
(self.share_inputs["draft_tokens"] if self.speculative_decoding else None),
self.share_inputs["seq_lens_encoder"],
@@ -864,7 +864,7 @@ class GPUModelRunner(ModelRunnerBase):
max_len_tensor_cpu=self.share_inputs["max_len_tensor_cpu"],
seq_lens_encoder=self.share_inputs["seq_lens_encoder"],
seq_lens_decoder=self.share_inputs["seq_lens_decoder"],
seq_lens_this_time=getattr(self.share_inputs, "seq_lens_this_time", self.seq_lens_this_time_buffer),
seq_lens_this_time=self.share_inputs["seq_lens_this_time"],
batch_id_per_token=self.share_inputs["batch_id_per_token"],
cu_seqlens_q=self.share_inputs["cu_seqlens_q"],
cu_seqlens_k=self.share_inputs["cu_seqlens_k"],