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 14, 05:30 UTC
[Misc] Avoid "PyTorch non-writable tensors" warning in RayPPCommunicator (#27443)
Signed-off-by: Rui Qiao <ruisearch42@gmail.com>
Rui Qiao committed
Oct 23, 2025 at 23:53 UTC
09a6a49eaf09c442c9c9c70a439dd892143c8be6
1 file changed
+1
-1
vllm/distributed/device_communicators/ray_communicator.py
+1
-1
@@ -99,7 +99,7 @@ class RayPPCommunicator(Communicator):
99
100
# Ray actor IDs are 32-character hex strings (128 bits)
101
ACTOR_ID_LEN = 32
102
- actor_id_bytes = actor_id_str.encode("utf-8")
102
+ actor_id_bytes = bytearray(actor_id_str.encode("utf-8"))
103
assert len(actor_id_bytes) == ACTOR_ID_LEN, (
104
f"Unexpected actor ID length: {len(actor_id_bytes)}"
105
)