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] disable processor cache (#19068)
Signed-off-by: raushan <raushan@huggingface.co>
Raushan Turganbay committed
Jun 3, 2025 at 17:06 UTC
d81edded69a5534a80785b68cde26c547cfcd4c6
1 file changed
+2
-2
vllm/v1/engine/mm_input_cache.py
+2
-2
@@ -34,8 +34,8 @@ class MirroredProcessingCache:
34
35
def __init__(self, model_config):
36
mm_config = model_config.multimodal_config
37
- disable_mm_preprocessor_cache = mm_config is not None and \
38
- not mm_config.disable_mm_preprocessor_cache
37
+ disable_mm_preprocessor_cache = (
38
+ mm_config is not None and mm_config.disable_mm_preprocessor_cache)
39
self.use_cache = not disable_mm_preprocessor_cache
40
self.mm_cache = ProcessingCache.get_lru_cache(VLLM_MM_INPUT_CACHE_GIB,
41
MultiModalKwargs)