fix: kimi k2 xgrammar crash (#8367)

Co-authored-by: cicirori <32845984+cicirori@users.noreply.github.com>
Co-authored-by: gongwei-130 <56567052+gongwei-130@users.noreply.github.com>
This commit is contained in:
Yineng Zhang
2025-07-25 15:44:01 -07:00
committed by GitHub
parent d8ee15643b
commit f8ca2368b2
2 changed files with 15 additions and 2 deletions

View File

@@ -1129,6 +1129,7 @@ class Scheduler(
bootstrap_port=recv_req.bootstrap_port,
bootstrap_room=recv_req.bootstrap_room,
data_parallel_rank=recv_req.data_parallel_rank,
vocab_size=self.model_config.vocab_size,
)
req.tokenizer = self.tokenizer
@@ -1395,8 +1396,10 @@ class Scheduler(
logger.info(f)
if self.enable_metrics:
cache_hit_rate = adder.log_hit_tokens / (
adder.log_input_tokens + adder.log_hit_tokens
total_tokens = adder.log_input_tokens + adder.log_hit_tokens
cache_hit_rate = (
adder.log_hit_tokens / total_tokens if total_tokens > 0 else 0.0
)
self.stats.num_running_reqs = running_bs
self.stats.num_used_tokens = num_used