[1/2] Support deterministic inference with flashinfer attention backend (#10645)

Co-authored-by: hebiao064 <hebiaobuaa@gmail.com> Co-authored-by: Qiaolin-Yu <liin1211@outlook.com>
2025-09-19 23:34:29 -07:00
parent 1d1ce62495
commit 8ecef73f12
10 changed files with 427 additions and 6 deletions
--- a/python/sglang/srt/sampling/sampling_batch_info.py
+++ b/python/sglang/srt/sampling/sampling_batch_info.py
@@ -75,6 +75,7 @@ class SamplingBatchInfo:
    @classmethod
    def from_schedule_batch(cls, batch: ScheduleBatch, vocab_size: int):
        global_server_args_dict = cls._get_global_server_args_dict()
+        enable_deterministic = global_server_args_dict["enable_deterministic_inference"]

        reqs = batch.reqs
        device = batch.device