{
  "id": 192015,
  "title": "Help looking for a certain scene",
  "url": "/competitions/lyft-motion-prediction-autonomous-vehicles/discussion/192015",
  "author_name": "",
  "post_date": "2020-10-19T20:19:23.975742Z",
  "votes": null,
  "comment_count": 3,
  "views": 0,
  "content": "<p>Hey guys,</p>\n<p>I'm not part of the competition but I'm trying to work with the dataset to get some data for my team's algorithm. Do you guys know if it's possible to look at a scene for an extended period of time? Like for example, I want to look at an intersection and see the scene play out for a minute or two even if the autonomous vehicle moves past the intersection. I would then extract extra info from the scene during that extended amount of time. Is this possible and if so, how would I go about approaching it? This is my first time using a dataset of any kind so I'm fairly new to it. Thank you.</p>",
  "messages": [
    {
      "id": "1054327",
      "postDate": "10/19/2020 20:19:23",
      "content": "<p>Hey guys,</p>\n<p>I'm not part of the competition but I'm trying to work with the dataset to get some data for my team's algorithm. Do you guys know if it's possible to look at a scene for an extended period of time? Like for example, I want to look at an intersection and see the scene play out for a minute or two even if the autonomous vehicle moves past the intersection. I would then extract extra info from the scene during that extended amount of time. Is this possible and if so, how would I go about approaching it? This is my first time using a dataset of any kind so I'm fairly new to it. Thank you.</p>",
      "rawMarkdown": "Hey guys,\n\nI'm not part of the competition but I'm trying to work with the dataset to get some data for my team's algorithm. Do you guys know if it's possible to look at a scene for an extended period of time? Like for example, I want to look at an intersection and see the scene play out for a minute or two even if the autonomous vehicle moves past the intersection. I would then extract extra info from the scene during that extended amount of time. Is this possible and if so, how would I go about approaching it? This is my first time using a dataset of any kind so I'm fairly new to it. Thank you.",
      "votes": null
    },
    {
      "id": "1054422",
      "postDate": "10/19/2020 23:21:57",
      "content": "<p>Hi <a href=\"https://www.kaggle.com/hushnudamridinov\" target=\"_blank\">@hushnudamridinov</a> Look at Output 23 from this file <a href=\"https://www.kaggle.com/corochann/lyft-comprehensive-guide-to-start-competition\" target=\"_blank\">https://www.kaggle.com/corochann/lyft-comprehensive-guide-to-start-competition</a> .<br>\nAre you looking for something like that? you can also refer to visualization notebook from l5kit repo. </p>",
      "rawMarkdown": "Hi @hushnudamridinov Look at Output 23 from this file https://www.kaggle.com/corochann/lyft-comprehensive-guide-to-start-competition .\nAre you looking for something like that? you can also refer to visualization notebook from l5kit repo.",
      "votes": null
    },
    {
      "id": "1054458",
      "postDate": "10/20/2020 00:25:12",
      "content": "<p>Hey <a href=\"https://www.kaggle.com/deepakrajpurushothaman\" target=\"_blank\">@deepakrajpurushothaman</a>. Yeah, I'm looking for something like that but I would like the scene to focus on an intersection instead of the AV so that I can observe what's happening at the intersection. Also, I was wondering if there was a way to access the speed of the AV and the vehicles around it. <br>\nI'm a bit familiar with the visualization notebook because that's mainly what my team is using to test out the dataset. </p>",
      "rawMarkdown": "Hey @deepakrajpurushothaman. Yeah, I'm looking for something like that but I would like the scene to focus on an intersection instead of the AV so that I can observe what's happening at the intersection. Also, I was wondering if there was a way to access the speed of the AV and the vehicles around it. \nI'm a bit familiar with the visualization notebook because that's mainly what my team is using to test out the dataset.",
      "votes": null
    },
    {
      "id": "1055499",
      "postDate": "10/20/2020 21:27:23",
      "content": "<p>Did you take a look into their official paper:</p>\n<p><a href=\"https://arxiv.org/pdf/2006.14480v1.pdf\" target=\"_blank\">https://arxiv.org/pdf/2006.14480v1.pdf</a></p>\n<p>It does answer a lot of questions, e.g. that each sequence is 25 seconds long so it wont be possible to play it for longer</p>",
      "rawMarkdown": "Did you take a look into their official paper:\n\nhttps://arxiv.org/pdf/2006.14480v1.pdf\n\nIt does answer a lot of questions, e.g. that each sequence is 25 seconds long so it wont be possible to play it for longer",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 1054422,
      "author_name": "deepakrajpurushothaman",
      "author_url": "",
      "post_date": "10/19/2020 23:21:57",
      "content": "<p>Hi <a href=\"https://www.kaggle.com/hushnudamridinov\" target=\"_blank\">@hushnudamridinov</a> Look at Output 23 from this file <a href=\"https://www.kaggle.com/corochann/lyft-comprehensive-guide-to-start-competition\" target=\"_blank\">https://www.kaggle.com/corochann/lyft-comprehensive-guide-to-start-competition</a> .<br>\nAre you looking for something like that? you can also refer to visualization notebook from l5kit repo. </p>",
      "votes": null,
      "replies": [
        {
          "id": 1054458,
          "author_name": "hushnudamridinov",
          "author_url": "",
          "post_date": "10/20/2020 00:25:12",
          "content": "<p>Hey <a href=\"https://www.kaggle.com/deepakrajpurushothaman\" target=\"_blank\">@deepakrajpurushothaman</a>. Yeah, I'm looking for something like that but I would like the scene to focus on an intersection instead of the AV so that I can observe what's happening at the intersection. Also, I was wondering if there was a way to access the speed of the AV and the vehicles around it. <br>\nI'm a bit familiar with the visualization notebook because that's mainly what my team is using to test out the dataset. </p>",
          "votes": null,
          "replies": []
        }
      ]
    },
    {
      "id": 1055499,
      "author_name": "aliabdin1",
      "author_url": "",
      "post_date": "10/20/2020 21:27:23",
      "content": "<p>Did you take a look into their official paper:</p>\n<p><a href=\"https://arxiv.org/pdf/2006.14480v1.pdf\" target=\"_blank\">https://arxiv.org/pdf/2006.14480v1.pdf</a></p>\n<p>It does answer a lot of questions, e.g. that each sequence is 25 seconds long so it wont be possible to play it for longer</p>",
      "votes": null,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "1054327": "Hey guys,\n\nI'm not part of the competition but I'm trying to work with the dataset to get some data for my team's algorithm. Do you guys know if it's possible to look at a scene for an extended period of time? Like for example, I want to look at an intersection and see the scene play out for a minute or two even if the autonomous vehicle moves past the intersection. I would then extract extra info from the scene during that extended amount of time. Is this possible and if so, how would I go about approaching it? This is my first time using a dataset of any kind so I'm fairly new to it. Thank you.",
    "1054422": "Hi @hushnudamridinov Look at Output 23 from this file https://www.kaggle.com/corochann/lyft-comprehensive-guide-to-start-competition .\nAre you looking for something like that? you can also refer to visualization notebook from l5kit repo.",
    "1054458": "Hey @deepakrajpurushothaman. Yeah, I'm looking for something like that but I would like the scene to focus on an intersection instead of the AV so that I can observe what's happening at the intersection. Also, I was wondering if there was a way to access the speed of the AV and the vehicles around it. \nI'm a bit familiar with the visualization notebook because that's mainly what my team is using to test out the dataset.",
    "1055499": "Did you take a look into their official paper:\n\nhttps://arxiv.org/pdf/2006.14480v1.pdf\n\nIt does answer a lot of questions, e.g. that each sequence is 25 seconds long so it wont be possible to play it for longer"
  },
  "source": "meta"
}