biondizzle/vllm
1
0
Fork 0
Code Issues Pull Requests Actions 2 Packages Projects Releases Wiki Activity

[Misc]Minor Changes about Worker (#11555)

Browse Source 
Signed-off-by: Chenguang Li <757486878@qq.com>
This commit is contained in:
Chenguang Li
2025-01-13 15:47:05 +08:00
committed by GitHub
parent cf6bbcb493
commit c3f05b09a0
2 changed files with 0 additions and 2 deletions
Download Patch File Download Diff File Expand all files Collapse all files

1
vllm/v1/worker/gpu_worker.py
View File

@@ -132,7 +132,6 @@ class Worker:
# Execute a forward pass with dummy inputs to profile the memory usage # Execute a forward pass with dummy inputs to profile the memory usage
# of the model. # of the model.
self.model_runner.profile_run() self.model_runner.profile_run()
torch.cuda.synchronize()
free_gpu_memory, _ = torch.cuda.mem_get_info() free_gpu_memory, _ = torch.cuda.mem_get_info()
# NOTE(woosuk): Here we assume that the other processes using the same # NOTE(woosuk): Here we assume that the other processes using the same

1
vllm/worker/worker.py
View File

@@ -200,7 +200,6 @@ class Worker(LocalOrDistributedWorkerBase):
weights_memory_in_bytes=self.model_runner. weights_memory_in_bytes=self.model_runner.
model_memory_usage) as result: model_memory_usage) as result:
self.model_runner.profile_run() self.model_runner.profile_run()
torch.cuda.synchronize()
self._assert_memory_footprint_increased_during_profiling() self._assert_memory_footprint_increased_during_profiling()