{
  "id": 191624,
  "title": "Can anyone fix this",
  "url": "/competitions/lyft-motion-prediction-autonomous-vehicles/discussion/191624",
  "author_name": "",
  "post_date": "2020-10-17T15:12:25.172754100Z",
  "votes": 4,
  "comment_count": 1,
  "views": 0,
  "content": "<p><img src=\"https://www.googleapis.com/download/storage/v1/b/kaggle-forum-message-attachments/o/inbox%2F3514793%2F12a88551660021aa54a74fb7606802fb%2FScreenshot%202020-10-17%20204020.jpg?generation=1602947470058915&amp;alt=media\" alt=\"\"></p>\n<h3>Why is this like this for checking I took <a href=\"https://www.kaggle.com/huanvo\" target=\"_blank\">@huanvo</a> high public kernel scoring weights and the result was above</h3>",
  "messages": [
    {
      "id": "1052294",
      "postDate": "10/17/2020 15:12:25",
      "content": "<p><img src=\"https://www.googleapis.com/download/storage/v1/b/kaggle-forum-message-attachments/o/inbox%2F3514793%2F12a88551660021aa54a74fb7606802fb%2FScreenshot%202020-10-17%20204020.jpg?generation=1602947470058915&amp;alt=media\" alt=\"\"></p>\n<h3>Why is this like this for checking I took <a href=\"https://www.kaggle.com/huanvo\" target=\"_blank\">@huanvo</a> high public kernel scoring weights and the result was above</h3>",
      "rawMarkdown": "![](https://www.googleapis.com/download/storage/v1/b/kaggle-forum-message-attachments/o/inbox%2F3514793%2F12a88551660021aa54a74fb7606802fb%2FScreenshot%202020-10-17%20204020.jpg?generation=1602947470058915&alt=media)\n\n### Why is this like this for checking I took @huanvo high public kernel scoring weights and the result was above",
      "votes": null
    },
    {
      "id": "1052579",
      "postDate": "10/18/2020 00:39:24",
      "content": "<p>sqrt(score / 25) gives an approximate Mae per step, it's about 2.2km <br>\nMaybe check you haven't swapped the x,y coordinates, and have the right localization centre for your predictions</p>\n<p>It wouldn't be non-local training targets? step 1 error &gt; step 50 </p>",
      "rawMarkdown": "sqrt(score / 25) gives an approximate Mae per step, it's about 2.2km \nMaybe check you haven't swapped the x,y coordinates, and have the right localization centre for your predictions\n\nIt wouldn't be non-local training targets? step 1 error > step 50",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 1052579,
      "author_name": "n3n77i",
      "author_url": "",
      "post_date": "10/18/2020 00:39:24",
      "content": "<p>sqrt(score / 25) gives an approximate Mae per step, it's about 2.2km <br>\nMaybe check you haven't swapped the x,y coordinates, and have the right localization centre for your predictions</p>\n<p>It wouldn't be non-local training targets? step 1 error &gt; step 50 </p>",
      "votes": null,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "1052294": "![](https://www.googleapis.com/download/storage/v1/b/kaggle-forum-message-attachments/o/inbox%2F3514793%2F12a88551660021aa54a74fb7606802fb%2FScreenshot%202020-10-17%20204020.jpg?generation=1602947470058915&alt=media)\n\n### Why is this like this for checking I took @huanvo high public kernel scoring weights and the result was above",
    "1052579": "sqrt(score / 25) gives an approximate Mae per step, it's about 2.2km \nMaybe check you haven't swapped the x,y coordinates, and have the right localization centre for your predictions\n\nIt wouldn't be non-local training targets? step 1 error > step 50"
  },
  "source": "meta"
}