xinyun/vllm
1
0
Fork 0
mirror of https://git.datalinker.icu/vllm-project/vllm.git synced 2026-09-09 09:47:02 +08:00
Code Issues Packages Projects Releases Wiki Activity

yapf

Browse Source
This commit is contained in:
Woosuk Kwon 2024-04-01 07:07:57 +00:00
parent 4cdb732cef
commit 31d05f7edb
1 changed files with 4 additions and 4 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

8
vllm/executor/tpu_executor.py
View File

@ -65,10 +65,10 @@ class TPUExecutor(ExecutorBase):
""" """
# Get the maximum number of blocks that can be allocated on TPU. # Get the maximum number of blocks that can be allocated on TPU.
num_tpu_blocks = self.driver_worker.profile_num_available_blocks( num_tpu_blocks = self.driver_worker.profile_num_available_blocks(
block_size=self.cache_config.block_size, block_size=self.cache_config.block_size,
gpu_memory_utilization=self.cache_config.gpu_memory_utilization, gpu_memory_utilization=self.cache_config.gpu_memory_utilization,
cache_dtype=self.cache_config.cache_dtype, cache_dtype=self.cache_config.cache_dtype,
) )
logger.info(f"# TPU blocks: {num_tpu_blocks}") logger.info(f"# TPU blocks: {num_tpu_blocks}")
check_block_size_valid(num_tpu_blocks, self.cache_config.block_size, check_block_size_valid(num_tpu_blocks, self.cache_config.block_size,