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
[Log] Debug Once for Randomizing dummy data for DP Rank (#22860)
Signed-off-by: yewentao256 <zhyanwentao@126.com>
Wentao Ye committed
Aug 15, 2025 at 14:51 UTC
df5afa82e5c41be7d87ddd1968e13891d22003a7
1 file changed
+1
-1
vllm/v1/worker/gpu_model_runner.py
+1
-1
@@ -2192,7 +2192,7 @@ class GPUModelRunner(LoRAModelRunnerMixin, KVConnectorModelRunnerMixin):
2192
high=self.model_config.get_vocab_size(),
2193
dtype=input_ids.dtype)
2194
2195
- logger.debug("Randomizing dummy data for DP Rank")
2195
+ logger.debug_once("Randomizing dummy data for DP Rank")
2196
input_ids.copy_(rand_input_ids()[:input_ids.size(0)],
2197
non_blocking=True)
2198
yield