xinyun/vllm
1
0
Fork 0
mirror of https://git.datalinker.icu/vllm-project/vllm.git synced 2025-12-10 08:04:58 +08:00
Code Issues Packages Projects Releases Wiki Activity

[XPU] fix xpu to set cudagraph batch sizes (#23044)

Browse Source 
Signed-off-by: calvin chen <wen.chen@dynamia.ai>
This commit is contained in:
Calvin Chen 2025-08-18 05:45:42 +08:00 committed by GitHub
parent 6d243efeda
commit 21e39436c8
No known key found for this signature in database
GPG Key ID: B5690EEEBB952194
1 changed files with 4 additions and 2 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

6
vllm/v1/worker/gpu_model_runner.py
View File

@ -232,8 +232,10 @@ class GPUModelRunner(LoRAModelRunnerMixin, KVConnectorModelRunnerMixin):
# The convention is different.
# self.cudagraph_batch_sizes sorts in ascending order.
# The batch sizes in the config are in descending order.
self.cudagraph_batch_sizes = list(
reversed(self.compilation_config.cudagraph_capture_sizes))
if self.compilation_config.cudagraph_capture_sizes and \
self.compilation_config.cudagraph_mode != CUDAGraphMode.NONE:
self.cudagraph_batch_sizes = list(
reversed(self.compilation_config.cudagraph_capture_sizes))
# Cache the device properties.
self._init_device_properties()