{
  "id": 37203,
  "title": "how to speed up csv generation?",
  "url": "/competitions/carvana-image-masking-challenge/discussion/37203",
  "author_name": "",
  "post_date": "2017-07-29T01:59:39.566186800Z",
  "votes": 3,
  "comment_count": 1,
  "views": 0,
  "content": "<p>I use CNN to make the mask of each test images. This is quite fast. \nHowever, I have problem in converting these mask to run length encoding and put in csv file.\nEach image is 1280x1918, and there are 100064 test images. It will take me several hours to make the csv file.</p>\n\n<p>How long do you guy takes to generate a csvfile? Anyone has sample code for multi-threading in python?</p>",
  "messages": [
    {
      "id": "208200",
      "postDate": "07/29/2017 01:59:39",
      "content": "<p>I use CNN to make the mask of each test images. This is quite fast. \nHowever, I have problem in converting these mask to run length encoding and put in csv file.\nEach image is 1280x1918, and there are 100064 test images. It will take me several hours to make the csv file.</p>\n\n<p>How long do you guy takes to generate a csvfile? Anyone has sample code for multi-threading in python?</p>",
      "rawMarkdown": "I use CNN to make the mask of each test images. This is quite fast. \nHowever, I have problem in converting these mask to run length encoding and put in csv file.\nEach image is 1280x1918, and there are 100064 test images. It will take me several hours to make the csv file.\n\nHow long do you guy takes to generate a csvfile? Anyone has sample code for multi-threading in python?",
      "votes": null
    },
    {
      "id": "208202",
      "postDate": "07/29/2017 02:07:15",
      "content": "<p>this code solve my problem!</p>\n\n<p><a href=\"https://www.kaggle.com/stainsby/fast-tested-rle\">https://www.kaggle.com/stainsby/fast-tested-rle</a></p>\n\n<p>thanks to @Sam StainsbyFast </p>",
      "rawMarkdown": "this code solve my problem!\n\nhttps://www.kaggle.com/stainsby/fast-tested-rle\n\nthanks to @Sam StainsbyFast",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 208202,
      "author_name": "hengck23",
      "author_url": "",
      "post_date": "07/29/2017 02:07:15",
      "content": "<p>this code solve my problem!</p>\n\n<p><a href=\"https://www.kaggle.com/stainsby/fast-tested-rle\">https://www.kaggle.com/stainsby/fast-tested-rle</a></p>\n\n<p>thanks to @Sam StainsbyFast </p>",
      "votes": null,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "208200": "I use CNN to make the mask of each test images. This is quite fast. \nHowever, I have problem in converting these mask to run length encoding and put in csv file.\nEach image is 1280x1918, and there are 100064 test images. It will take me several hours to make the csv file.\n\nHow long do you guy takes to generate a csvfile? Anyone has sample code for multi-threading in python?",
    "208202": "this code solve my problem!\n\nhttps://www.kaggle.com/stainsby/fast-tested-rle\n\nthanks to @Sam StainsbyFast"
  },
  "source": "meta"
}