{
  "id": 390296,
  "title": "Input preprocessing before prediction?",
  "url": "/competitions/asl-signs/discussion/390296",
  "author_name": "",
  "post_date": "2023-02-25T00:31:14.772445Z",
  "votes": null,
  "comment_count": 3,
  "views": 0,
  "content": "<p>Can there be a data-preprocessing function right before the call to the model's prediction? </p>\n<p>The Evaluation indicates that there will be a \"load_relevant_data_subset\" function, which returns a ndarray.shape(n, 543,3).  What if I wanted to take that ndarray and reshape it or split it in batches to format the input in a certain way to feed my model?</p>\n<p>Thanks.</p>",
  "messages": [
    {
      "id": "2158547",
      "postDate": "02/25/2023 00:31:14",
      "content": "<p>Can there be a data-preprocessing function right before the call to the model's prediction? </p>\n<p>The Evaluation indicates that there will be a \"load_relevant_data_subset\" function, which returns a ndarray.shape(n, 543,3).  What if I wanted to take that ndarray and reshape it or split it in batches to format the input in a certain way to feed my model?</p>\n<p>Thanks.</p>",
      "rawMarkdown": "Can there be a data-preprocessing function right before the call to the model's prediction? \n\nThe Evaluation indicates that there will be a \"load_relevant_data_subset\" function, which returns a ndarray.shape(n, 543,3).  What if I wanted to take that ndarray and reshape it or split it in batches to format the input in a certain way to feed my model?\n\nThanks.",
      "votes": null
    },
    {
      "id": "2158550",
      "postDate": "02/25/2023 00:36:56",
      "content": "<p>I think you can do that inside your model </p>",
      "rawMarkdown": "I think you can do that inside your model",
      "votes": null
    },
    {
      "id": "2158626",
      "postDate": "02/25/2023 02:56:15",
      "content": "<p>Thank you!</p>",
      "rawMarkdown": "Thank you!",
      "votes": null
    },
    {
      "id": "2160181",
      "postDate": "02/26/2023 12:57:38",
      "content": "<p>Can anyone point us to documentation that outlines what you can and can not do in tfLite? By writing custom layers in Tf, you can indeed do many things, but in order to focus our effort, it would be good to know about the restrictions in advance.</p>",
      "rawMarkdown": "Can anyone point us to documentation that outlines what you can and can not do in tfLite? By writing custom layers in Tf, you can indeed do many things, but in order to focus our effort, it would be good to know about the restrictions in advance.",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 2158550,
      "author_name": "camryanwashere",
      "author_url": "",
      "post_date": "02/25/2023 00:36:56",
      "content": "<p>I think you can do that inside your model </p>",
      "votes": null,
      "replies": [
        {
          "id": 2158626,
          "author_name": "juanolano",
          "author_url": "",
          "post_date": "02/25/2023 02:56:15",
          "content": "<p>Thank you!</p>",
          "votes": null,
          "replies": []
        },
        {
          "id": 2160181,
          "author_name": "wonderingalice",
          "author_url": "",
          "post_date": "02/26/2023 12:57:38",
          "content": "<p>Can anyone point us to documentation that outlines what you can and can not do in tfLite? By writing custom layers in Tf, you can indeed do many things, but in order to focus our effort, it would be good to know about the restrictions in advance.</p>",
          "votes": null,
          "replies": []
        }
      ]
    }
  ],
  "raw_markdown_by_id": {
    "2158547": "Can there be a data-preprocessing function right before the call to the model's prediction? \n\nThe Evaluation indicates that there will be a \"load_relevant_data_subset\" function, which returns a ndarray.shape(n, 543,3).  What if I wanted to take that ndarray and reshape it or split it in batches to format the input in a certain way to feed my model?\n\nThanks.",
    "2158550": "I think you can do that inside your model",
    "2158626": "Thank you!",
    "2160181": "Can anyone point us to documentation that outlines what you can and can not do in tfLite? By writing custom layers in Tf, you can indeed do many things, but in order to focus our effort, it would be good to know about the restrictions in advance."
  },
  "source": "meta"
}