A high-throughput and memory-efficient inference and serving engine for LLMs
Mirrored from https://github.com/tduncan-lum/Nvidia-Optimization-Stack---VLLM.git
Synced Nov 17, 00:30 UTC
[misc][typo] fix typo (#5372)
youkaichao committed
Jun 10, 2024 at 02:51 UTC
c81da5f56dd613d5378078866e4810452d50d6ed
1 file changed
+1
-1
vllm/distributed/device_communicators/custom_all_reduce_utils.py
+1
-1
@@ -166,7 +166,7 @@ def gpu_p2p_access_check(i: int, j: int) -> bool:
166
and (not os.path.exists(path))):
167
# only the local master process (with local_rank == 0) can
168
# enter this block to calculate the cache
169
- logger.info("generating GPU P2P access cache for in %s", path)
169
+ logger.info("generating GPU P2P access cache in %s", path)
170
cache = {}
171
for _i in range(num_dev):
172
for _j in range(num_dev):