{
  "id": 415563,
  "title": "Problem on Matcha model",
  "url": "/competitions/benetech-making-graphs-accessible/discussion/415563",
  "author_name": "",
  "post_date": "2023-06-07T03:17:54.787544500Z",
  "votes": 1,
  "comment_count": 2,
  "views": 0,
  "content": "<p>I used the donut model and it didnot work well. I tried to work on the Matcha model and finetune it. I modify the dataset class and collator function, but something go wrong. It raises an error on TypeError: Pix2StructImageProcessor object got multiple values for keyword argument 'header_text'. <br>\nIt really take me huge of time, and I donnot know why. Hope that anybody can figure it out. Thanks a lot!<br>\nHere is the error <a href=\"https://www.kaggle.com/code/yyewdy/training-matcha-error\" target=\"_blank\">notebook</a></p>",
  "messages": [
    {
      "id": "2290672",
      "postDate": "06/07/2023 03:17:54",
      "content": "<p>I used the donut model and it didnot work well. I tried to work on the Matcha model and finetune it. I modify the dataset class and collator function, but something go wrong. It raises an error on TypeError: Pix2StructImageProcessor object got multiple values for keyword argument 'header_text'. <br>\nIt really take me huge of time, and I donnot know why. Hope that anybody can figure it out. Thanks a lot!<br>\nHere is the error <a href=\"https://www.kaggle.com/code/yyewdy/training-matcha-error\" target=\"_blank\">notebook</a></p>",
      "rawMarkdown": "I used the donut model and it didnot work well. I tried to work on the Matcha model and finetune it. I modify the dataset class and collator function, but something go wrong. It raises an error on TypeError: Pix2StructImageProcessor object got multiple values for keyword argument 'header_text'. \nIt really take me huge of time, and I donnot know why. Hope that anybody can figure it out. Thanks a lot!\nHere is the error [notebook](https://www.kaggle.com/code/yyewdy/training-matcha-error)",
      "votes": null
    },
    {
      "id": "2290817",
      "postDate": "06/07/2023 06:03:52",
      "content": "<p><code>processor = AutoProcessor.from_pretrained(Config['MODEL_NAME'],is_vqa=False)\n</code><br>\nHi,<a href=\"https://www.kaggle.com/yyewdy\" target=\"_blank\">@yyewdy</a>,you can try this line of code</p>",
      "rawMarkdown": "`    processor = AutoProcessor.from_pretrained(Config['MODEL_NAME'],is_vqa=False)\n`\nHi,[@yyewdy](https://www.kaggle.com/yyewdy),you can try this line of code",
      "votes": null
    },
    {
      "id": "2291484",
      "postDate": "06/07/2023 15:21:10",
      "content": "<p>Thanks a lot! I think I fix the problem!</p>",
      "rawMarkdown": "Thanks a lot! I think I fix the problem!",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 2290817,
      "author_name": "areno9516",
      "author_url": "",
      "post_date": "06/07/2023 06:03:52",
      "content": "<p><code>processor = AutoProcessor.from_pretrained(Config['MODEL_NAME'],is_vqa=False)\n</code><br>\nHi,<a href=\"https://www.kaggle.com/yyewdy\" target=\"_blank\">@yyewdy</a>,you can try this line of code</p>",
      "votes": null,
      "replies": [
        {
          "id": 2291484,
          "author_name": "yyewdy",
          "author_url": "",
          "post_date": "06/07/2023 15:21:10",
          "content": "<p>Thanks a lot! I think I fix the problem!</p>",
          "votes": null,
          "replies": []
        }
      ]
    }
  ],
  "raw_markdown_by_id": {
    "2290672": "I used the donut model and it didnot work well. I tried to work on the Matcha model and finetune it. I modify the dataset class and collator function, but something go wrong. It raises an error on TypeError: Pix2StructImageProcessor object got multiple values for keyword argument 'header_text'. \nIt really take me huge of time, and I donnot know why. Hope that anybody can figure it out. Thanks a lot!\nHere is the error [notebook](https://www.kaggle.com/code/yyewdy/training-matcha-error)",
    "2290817": "`    processor = AutoProcessor.from_pretrained(Config['MODEL_NAME'],is_vqa=False)\n`\nHi,[@yyewdy](https://www.kaggle.com/yyewdy),you can try this line of code",
    "2291484": "Thanks a lot! I think I fix the problem!"
  },
  "source": "meta"
}