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, 07:15 UTC
[Bugfix]: Fix TypeError: 'float' object cannot be interpreted as an integer (#19283)
Signed-off-by: chaunceyjiang <chaunceyjiang@gmail.com>
Chauncey committed
Jun 8, 2025 at 08:16 UTC
d77f7fb871473863df48ea71b4a8706028048f19
1 file changed
+1
-1
vllm/v1/utils.py
+1
-1
@@ -407,7 +407,7 @@ class CoreEngineActorManager:
407
# For now, each DP rank can only be assigned to one node
408
# TODO(rui): support allocating a single DP rank
409
# to multiple nodes
410
- available_engine_count = node_resources["GPU"] // world_size
410
+ available_engine_count = int(node_resources["GPU"]) // world_size
411
if node_ip == dp_master_ip:
412
assert available_engine_count >= local_engine_count, (
413
"Not enough resources to allocate DP ranks "