[V0 Deprecation] Remove VLLM_USE_V1 from tests (#26341)

Signed-off-by: DarkLight1337 <tlleungac@connect.ust.hk>
This commit is contained in:
Cyrus Leung
2025-10-07 15:42:31 +00:00
committed by GitHub
parent c0a7b89d8e
commit 1e4ecca1d0
51 changed files with 816 additions and 1274 deletions
+7 -1
View File
@@ -77,7 +77,13 @@ class CPUModelRunner(GPUModelRunner):
logger.info("Warming up model for the compilation...")
# Only generate graph for the generic shape
with _set_global_compilation_settings(self.vllm_config):
self._dummy_run(max(16, self.max_num_reqs))
self._dummy_run(
min(
max(16, self.max_num_reqs),
self.scheduler_config.max_num_batched_tokens,
)
)
logger.info("Warming up done.")
def _init_device_properties(self) -> None: