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 7, 21:30 UTC
gpu_model_runner: Cache is_encoder_decoder from model config (#35099)
Signed-off-by: Patrick Schlangen <pschlan@amd.com>
pschlan-amd committed
Feb 24, 2026 at 04:08 UTC
80d93fd6daf60d497c55a09c6dcd5471081c5978
1 file changed
+1
-1
vllm/config/model.py
+1
-1
@@ -1365,7 +1365,7 @@ class ModelConfig:
1365
1366
return diff_sampling_param
1367
1368
- @property
1368
+ @cached_property
1369
def is_encoder_decoder(self) -> bool:
1370
"""Extract the HF encoder/decoder model flag."""
1371
return is_encoder_decoder(self.hf_config)