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, 00:45 UTC
[misc] Revert # 12833 (#12857)
Signed-off-by: <> Co-authored-by: EC2 Default User <ec2-user@ip-172-31-20-117.us-west-2.compute.internal>
Kevin H. Luu committed
Feb 6, 2025 at 16:29 UTC
b26078235722a434d92fe90dcea2023a5ae7294a
1 file changed
+3
vllm/inputs/preprocess.py
+3
@@ -260,6 +260,9 @@ class InputPreprocessor:
260
mm_processor = self.mm_registry.create_processor(
261
self.model_config, tokenizer)
262
263
+ if isinstance(prompt, list):
264
+ prompt = tokenizer.decode(prompt)
265
+
266
if mm_processor_kwargs is None:
267
mm_processor_kwargs = {}
268