feat: return partial generation results when aborting requests in waiting queue (#11673)

This commit is contained in:
Yuhong Guo
2025-10-29 22:03:00 +08:00
committed by GitHub
parent 750940ae36
commit caa5d2967c
6 changed files with 168 additions and 52 deletions

View File

@@ -135,6 +135,7 @@ class Envs:
# Scheduler: memory leak test
SGLANG_TEST_RETRACT = EnvBool(False)
SGLANG_TEST_RETRACT_INTERVAL = EnvInt(3)
SGLANG_TEST_RETRACT_NO_PREFILL_BS = EnvInt(2 ** 31)
SGLANG_ENABLE_RUNTIME_MEM_LEAK_CHECK = EnvBool(False)
# Scheduler: new token ratio hyperparameters