{
  "id": 63559,
  "title": "how to generate prediction string",
  "url": "/competitions/google-ai-open-images-visual-relationship-track/discussion/63559",
  "author_name": "",
  "post_date": "2018-08-18T07:33:50.893106500Z",
  "votes": null,
  "comment_count": 1,
  "views": 0,
  "content": "<p>Can anyone help me in generating the prediction string.</p>",
  "messages": [
    {
      "id": "372068",
      "postDate": "08/18/2018 07:33:50",
      "content": "<p>Can anyone help me in generating the prediction string.</p>",
      "rawMarkdown": "Can anyone help me in generating the prediction string.",
      "votes": null
    },
    {
      "id": "372483",
      "postDate": "08/19/2018 14:00:26",
      "content": "<p>It will depend on the way you get your predictions out, if you can give me an example of how you get your predictions output, I can help you generate a prediction string.\nWhat I do is, if you know pandas its easy. I have the bounding box values for a single image as a dataframe then after I get the VR predictions, I append it to the dataframe and convert it to a list and remove the strings stuffs like commas and other things and mine works in a loop so for every image I create a new df and then do the above processing and then combine the ImageId with the prediction string and clear the old df.\nThe way my model works, this is kinda the efficient way I was able to find, there may be others but it depends on how you get your predictions</p>",
      "rawMarkdown": "It will depend on the way you get your predictions out, if you can give me an example of how you get your predictions output, I can help you generate a prediction string.\nWhat I do is, if you know pandas its easy. I have the bounding box values for a single image as a dataframe then after I get the VR predictions, I append it to the dataframe and convert it to a list and remove the strings stuffs like commas and other things and mine works in a loop so for every image I create a new df and then do the above processing and then combine the ImageId with the prediction string and clear the old df.\nThe way my model works, this is kinda the efficient way I was able to find, there may be others but it depends on how you get your predictions",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 372483,
      "author_name": "mukeshmithrakumar",
      "author_url": "",
      "post_date": "08/19/2018 14:00:26",
      "content": "<p>It will depend on the way you get your predictions out, if you can give me an example of how you get your predictions output, I can help you generate a prediction string.\nWhat I do is, if you know pandas its easy. I have the bounding box values for a single image as a dataframe then after I get the VR predictions, I append it to the dataframe and convert it to a list and remove the strings stuffs like commas and other things and mine works in a loop so for every image I create a new df and then do the above processing and then combine the ImageId with the prediction string and clear the old df.\nThe way my model works, this is kinda the efficient way I was able to find, there may be others but it depends on how you get your predictions</p>",
      "votes": null,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "372068": "Can anyone help me in generating the prediction string.",
    "372483": "It will depend on the way you get your predictions out, if you can give me an example of how you get your predictions output, I can help you generate a prediction string.\nWhat I do is, if you know pandas its easy. I have the bounding box values for a single image as a dataframe then after I get the VR predictions, I append it to the dataframe and convert it to a list and remove the strings stuffs like commas and other things and mine works in a loop so for every image I create a new df and then do the above processing and then combine the ImageId with the prediction string and clear the old df.\nThe way my model works, this is kinda the efficient way I was able to find, there may be others but it depends on how you get your predictions"
  },
  "source": "meta"
}