apps/FastDeploy
1
0
Fork 0
mirror of https://github.com/PaddlePaddle/FastDeploy.git synced 2025-10-04 16:22:57 +08:00
Code Issues Actions 2 Packages Projects Releases Wiki Activity

[XPU] Remove padding_offsets from get_padding_offset.cu (#2911)

Browse Source
This commit is contained in:
周周周
2025-07-18 14:16:44 +08:00
committed by GitHub
parent 0eb5dc18d3
commit 1339e56282
5 changed files with 24 additions and 24 deletions
Download Patch File Download Diff File Expand all files Collapse all files

6
fastdeploy/worker/xpu_model_runner.py
View File

@@ -58,7 +58,7 @@ def xpu_pre_process(
(
ids_remove_padding,
cum_offsets,
padding_offset,
batch_id_per_token,
cu_seqlens_q,
cu_seqlens_k,
) = get_padding_offset(input_ids, cum_offsets_now, token_num,
@@ -66,7 +66,7 @@ def xpu_pre_process(
share_inputs["ids_remove_padding"] = None # set this after adjust batch
share_inputs["cum_offsets"] = cum_offsets
share_inputs["padding_offset"] = padding_offset
share_inputs["batch_id_per_token"] = batch_id_per_token
share_inputs["cu_seqlens_q"] = cu_seqlens_q
share_inputs["cu_seqlens_k"] = cu_seqlens_k
@@ -79,7 +79,7 @@ def xpu_pre_process(
seq_lens_decoder=share_inputs["seq_lens_decoder"],
seq_lens_this_time=share_inputs["seq_lens_this_time"],
cum_offsets=share_inputs["cum_offsets"],
padding_offset=share_inputs["padding_offset"],
batch_id_per_token=share_inputs["batch_id_per_token"],
cu_seqlens_q=share_inputs["cu_seqlens_q"],
cu_seqlens_k=share_inputs["cu_seqlens_k"],
block_tables=share_inputs["block_tables"],