{
  "id": 410570,
  "title": "Separating the dataset - Second Fragment is really that BIG!",
  "url": "/competitions/vesuvius-challenge-ink-detection/discussion/410570",
  "author_name": "",
  "post_date": "2023-05-15T19:55:31.322692200Z",
  "votes": null,
  "comment_count": 1,
  "views": 0,
  "content": "<p><strong>Hello there, I'm having an ask, could anyone please upload the second fragment as a dataset to be able to train the model on it!! the 2 fragments need around 18G ram which is really huge while using GPU, and for me, \nit's too tough for me to download the dataset, so if there is someone who already downloaded the data then it's half of the way. just cut the image stack in half and upload it here.  As I think that its the most important and denser fragment here</strong><br>\n thanks.</p>",
  "messages": [
    {
      "id": "2260684",
      "postDate": "05/15/2023 19:55:31",
      "content": "<p><strong>Hello there, I'm having an ask, could anyone please upload the second fragment as a dataset to be able to train the model on it!! the 2 fragments need around 18G ram which is really huge while using GPU, and for me, \nit's too tough for me to download the dataset, so if there is someone who already downloaded the data then it's half of the way. just cut the image stack in half and upload it here.  As I think that its the most important and denser fragment here</strong><br>\n thanks.</p>",
      "rawMarkdown": "**Hello there, I'm having an ask, could anyone please upload the second fragment as a dataset to be able to train the model on it!! the 2 fragments need around 18G ram which is really huge while using GPU, and for me, \nit's too tough for me to download the dataset, so if there is someone who already downloaded the data then it's half of the way. just cut the image stack in half and upload it here.  As I think that its the most important and denser fragment here**\n thanks.",
      "votes": null
    },
    {
      "id": "2260858",
      "postDate": "05/16/2023 00:11:15",
      "content": "<p>either pre process to png 8 bits (or just slice the wanted fragment…) on a notebook and zip it on kaggle before downloading the ouput , or just get  a dataset pre processed like <a href=\"https://www.kaggle.com/datasets/tmyok1984/vcid-tile-images\" target=\"_blank\">here</a></p>",
      "rawMarkdown": "either pre process to png 8 bits (or just slice the wanted fragment...) on a notebook and zip it on kaggle before downloading the ouput , or just get  a dataset pre processed like [here](https://www.kaggle.com/datasets/tmyok1984/vcid-tile-images)",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 2260858,
      "author_name": "iraqbot",
      "author_url": "",
      "post_date": "05/16/2023 00:11:15",
      "content": "<p>either pre process to png 8 bits (or just slice the wanted fragment…) on a notebook and zip it on kaggle before downloading the ouput , or just get  a dataset pre processed like <a href=\"https://www.kaggle.com/datasets/tmyok1984/vcid-tile-images\" target=\"_blank\">here</a></p>",
      "votes": null,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "2260684": "**Hello there, I'm having an ask, could anyone please upload the second fragment as a dataset to be able to train the model on it!! the 2 fragments need around 18G ram which is really huge while using GPU, and for me, \nit's too tough for me to download the dataset, so if there is someone who already downloaded the data then it's half of the way. just cut the image stack in half and upload it here.  As I think that its the most important and denser fragment here**\n thanks.",
    "2260858": "either pre process to png 8 bits (or just slice the wanted fragment...) on a notebook and zip it on kaggle before downloading the ouput , or just get  a dataset pre processed like [here](https://www.kaggle.com/datasets/tmyok1984/vcid-tile-images)"
  },
  "source": "meta"
}