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 16, 05:45 UTC
[torch.compile] fix deprecated code (#10948)
Signed-off-by: youkaichao <youkaichao@gmail.com>
youkaichao committed
Dec 6, 2024 at 03:01 UTC
a1887f2c96480e597db8c35cb8389c4025fb4db9
1 file changed
+1
-1
vllm/compilation/backends.py
+1
-1
@@ -40,7 +40,7 @@ def wrap_inductor(graph,
40
compilation_counter.num_inductor_compilations += 1
41
42
from torch._inductor import config
43
- current_config = config.shallow_copy_dict()
43
+ current_config = config.get_config_copy()
44
from torch._inductor.compile_fx import compile_fx
45
46
if additional_inductor_config is not None: