@tduncan / Nvidia-Optimization-Stack-VLLM Mirror · read-only / commits / 7f60520deb
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, 10:45 UTC

[Misc] Update Default Image Mapper Error Log (#8977)

Signed-off-by: Alex-Brooks <Alex.Brooks@ibm.com> Co-authored-by: Roger Wang <136131678+ywang96@users.noreply.github.com>

Alex Brooks committed Oct 2, 2024 at 05:44 UTC 7f60520deb05d2e097b408e3310f1d383fbf1de6
1 file changed +6 -1
vllm/multimodal/image.py
+6 -1
@@ -56,7 +56,12 @@ class ImagePlugin(MultiModalPlugin):
56 .preprocess(data, return_tensors="pt") \
57 .data
58 except Exception:
59 - logger.error("Failed to process image (%s)", data)
59 + logger.error(
60 + "Failed to process image (%s) with the default mapper. "
61 + "This is most likely an edge-case with this model's image "
62 + "processor in transformers (type: %s), and not vLLM.",
63 + data,
64 + type(image_processor).__name__)
65 raise
66
67 return MultiModalInputs(batch_data)