Fix a bug in BatchTokenIDOut & Misc style and dependency updates (#7457)

This commit is contained in:
Lianmin Zheng
2025-06-23 06:20:39 -07:00
committed by GitHub
parent 8aa68ed5c4
commit 55e03b10c4
9 changed files with 37 additions and 32 deletions

View File

@@ -1374,7 +1374,14 @@ class Scheduler(
)
raise ValueError(msg)
if len(self.req_to_token_pool.free_slots) != self.req_to_token_pool.size:
if self.disaggregation_mode == DisaggregationMode.DECODE:
req_total_size = (
self.req_to_token_pool.size + self.req_to_token_pool.pre_alloc_size
)
else:
req_total_size = self.req_to_token_pool.size
if len(self.req_to_token_pool.free_slots) != req_total_size:
msg = (
"req_to_token_pool memory leak detected!"
f"available_size={len(self.req_to_token_pool.free_slots)}, "