Yin Chen

[W:onnxruntime:, execution_frame.cc:770 VerifyOutputSizes]

1

# Ask a Question ### Question I have a onnx model converted from pth. I inference sucessfully on CPU with CPUExecutionProvider, but meet the warning "[W:onnxruntime:, execution_frame.cc:770 VerifyOutputSizes] Expected shape...

Cannot open display ''.

1

when i run the script, it hsow Cannot open display ''.

Are there any plans to support this in Obsidian in the future?

Can the inference code of LLAVA-OV be modified to support batch inference? Single-instance inference is a bit slow.

1

```python3 from operator import attrgetter from llava.model.builder import load_pretrained_model from llava.mm_utils import get_model_name_from_path, process_images, tokenizer_image_token from llava.constants import IMAGE_TOKEN_INDEX, DEFAULT_IMAGE_TOKEN, DEFAULT_IM_START_TOKEN, DEFAULT_IM_END_TOKEN, IGNORE_INDEX from llava.conversation import conv_templates, SeparatorStyle import torch...