apps/FastDeploy
1
0
Fork 0
mirror of https://github.com/PaddlePaddle/FastDeploy.git synced 2026-04-23 08:21:53 +08:00
Code Issues Actions 19 Packages Projects Releases Wiki Activity

check (#5736)

Browse Source
This commit is contained in:
chen
2025-12-24 16:49:20 +08:00
committed by GitHub
parent 6b0fba8294
commit c7ab32d154
1 changed files with 5 additions and 0 deletions
Download Patch File Download Diff File Expand all files Collapse all files
custom_ops/gpu_ops/update_inputs_v1.cu
+5
View File
@@ -50,6 +50,11 @@ __global__ void update_inputs_kernel_v1(bool* not_need_stop,
}
if (thread_idx < bsz) {
if (stop_flag_now) {
// chuned when max_tokens=1
if (seq_lens_this_time[thread_idx] + seq_lens_decoder[thread_idx] <
prompt_lens[thread_idx]) {
topk_ids[thread_idx] = -1;
}
seq_lens_this_time[thread_idx] = 0; // stop at next step
seq_lens_decoder[thread_idx] = 0;
seq_lens_encoder[thread_idx] = 0;