when loading model, error raised like ValueError: size must contain 'shortest_edge' and 'longest_edge' keys.

[wuutiing]

ValueError: size must contain 'shortest_edge' and 'longest_edge' keys.

related code is

https://github.com/Alpha-Innovator/OmniCaptioner/blob/4c07d277d02f326973da8d15c3625d526ed4ac82/src/inference_single_image.py#L191-L195

transformer version is 4.51.2

how to fix that?

[wuutiing]

seems edit preprocessor_config.json in https://huggingface.co/U4R/OmniCaptioner/tree/main will be ok

  "size": {
    "max_pixels": 5017600,
    "shortest_edge": 56,
    "min_pixels": 3136,
    "longest_edge": 2240
  },