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 15, 22:15 UTC
[Bugfix][CPU] Fix cpu all-reduce using native pytorch implementation (#13586)
Isotr0py committed
Feb 21, 2025 at 14:24 UTC
b2c3fc5d65485bfed796ee11c2a019d70357ad10
1 file changed
+2
-1
vllm/distributed/device_communicators/cpu_communicator.py
+2
-1
@@ -30,4 +30,5 @@ class CpuCommunicator(DeviceCommunicatorBase):
30
pass
31
32
def all_reduce(self, input_):
33
- return self.dist_module.all_reduce(input_, group=self.device_group)
33
+ self.dist_module.all_reduce(input_, group=self.device_group)
34
+ return input_