{
  "id": 351534,
  "title": "Is it possible to preprocess and save image and load that image during inferencing hidden dataset in kaggle?",
  "url": "/competitions/mayo-clinic-strip-ai/discussion/351534",
  "author_name": "kaggler",
  "post_date": "2022-09-10T16:32:25.905000",
  "votes": 3,
  "comment_count": 6,
  "views": 0,
  "content": "<p>I have a question for kagglers<br>\nIs it possible to preprocess and save an image and load that image while inferencing the hidden dataset in kaggle?<br>\nI have never tried it before, a help would be very appreciated!</p>",
  "messages": [
    {
      "id": 1933636,
      "postDate": "2022-09-10T16:32:25.907Z",
      "content": "<p>I have a question for kagglers<br>\nIs it possible to preprocess and save an image and load that image while inferencing the hidden dataset in kaggle?<br>\nI have never tried it before, a help would be very appreciated!</p>",
      "rawMarkdown": "I have a question for kagglers\nIs it possible to preprocess and save an image and load that image while inferencing the hidden dataset in kaggle?\nI have never tried it before, a help would be very appreciated!",
      "votes": 3
    },
    {
      "id": 1934430,
      "postDate": "2022-09-11T11:12:27.757Z",
      "content": "<p>I think its possible but not efficient, in inference, you pre-process the image and compute the probability… add more steps slow down the execution.</p>",
      "rawMarkdown": "I think its possible but not efficient, in inference, you pre-process the image and compute the probability... add more steps slow down the execution.",
      "votes": 2,
      "replies": [
        {
          "id": 1969805,
          "postDate": "2022-10-03T17:35:50.970Z",
          "content": "<p>Very thanks for the reply! good luck with your position!</p>",
          "rawMarkdown": "Very thanks for the reply! good luck with your position!"
        }
      ]
    },
    {
      "id": 1953616,
      "postDate": "2022-09-24T15:59:19.097Z",
      "content": "<p>It's possible but why would you save it and load it again? It would make your inference time way longer than it should be.</p>",
      "rawMarkdown": "It's possible but why would you save it and load it again? It would make your inference time way longer than it should be.",
      "votes": 1,
      "replies": [
        {
          "id": 1969803,
          "postDate": "2022-10-03T17:35:35.143Z",
          "content": "<p>Thanks for the reply!</p>",
          "rawMarkdown": "Thanks for the reply!"
        }
      ]
    },
    {
      "id": 1946162,
      "postDate": "2022-09-19T16:28:06.617Z",
      "content": "<p>May be a good idea if your code cannot handle memory during preprocessing and inference together. Maybe viable if preprocessed image is saved for a fraction of images that are very large. Will need to waste one (or more) submissions to check it though :)</p>",
      "rawMarkdown": "May be a good idea if your code cannot handle memory during preprocessing and inference together. Maybe viable if preprocessed image is saved for a fraction of images that are very large. Will need to waste one (or more) submissions to check it though :)",
      "votes": 1,
      "replies": [
        {
          "id": 1953533,
          "postDate": "2022-09-24T14:45:53.883Z",
          "content": "<p>Thanks for the reply!</p>",
          "rawMarkdown": "Thanks for the reply!"
        }
      ]
    }
  ],
  "comments": [
    {
      "id": 1934430,
      "author_name": "felipe higon",
      "author_url": "",
      "post_date": "2022-09-11T11:12:27.757000",
      "content": "<p>I think its possible but not efficient, in inference, you pre-process the image and compute the probability… add more steps slow down the execution.</p>",
      "votes": 2,
      "replies": [
        {
          "id": 1969805,
          "author_name": "kaggler",
          "author_url": "",
          "post_date": "2022-10-03T17:35:50.970000",
          "content": "<p>Very thanks for the reply! good luck with your position!</p>",
          "votes": 0,
          "replies": []
        }
      ]
    },
    {
      "id": 1953616,
      "author_name": "Gunes Evitan",
      "author_url": "",
      "post_date": "2022-09-24T15:59:19.097000",
      "content": "<p>It's possible but why would you save it and load it again? It would make your inference time way longer than it should be.</p>",
      "votes": 1,
      "replies": [
        {
          "id": 1969803,
          "author_name": "kaggler",
          "author_url": "",
          "post_date": "2022-10-03T17:35:35.143000",
          "content": "<p>Thanks for the reply!</p>",
          "votes": 0,
          "replies": []
        }
      ]
    },
    {
      "id": 1946162,
      "author_name": "tdiceman",
      "author_url": "",
      "post_date": "2022-09-19T16:28:06.617000",
      "content": "<p>May be a good idea if your code cannot handle memory during preprocessing and inference together. Maybe viable if preprocessed image is saved for a fraction of images that are very large. Will need to waste one (or more) submissions to check it though :)</p>",
      "votes": 1,
      "replies": [
        {
          "id": 1953533,
          "author_name": "kaggler",
          "author_url": "",
          "post_date": "2022-09-24T14:45:53.883000",
          "content": "<p>Thanks for the reply!</p>",
          "votes": 0,
          "replies": []
        }
      ]
    }
  ],
  "raw_markdown_by_id": {
    "1933636": "I have a question for kagglers\nIs it possible to preprocess and save an image and load that image while inferencing the hidden dataset in kaggle?\nI have never tried it before, a help would be very appreciated!",
    "1934430": "I think its possible but not efficient, in inference, you pre-process the image and compute the probability... add more steps slow down the execution.",
    "1953616": "It's possible but why would you save it and load it again? It would make your inference time way longer than it should be.",
    "1946162": "May be a good idea if your code cannot handle memory during preprocessing and inference together. Maybe viable if preprocessed image is saved for a fraction of images that are very large. Will need to waste one (or more) submissions to check it though :)"
  }
}