@tduncan / Nvidia-Optimization-Stack-VLLM Mirror · read-only / commits / b260782357
Fork 0

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