{
  "id": 193159,
  "title": "Model input and output sizes  ",
  "url": "/competitions/lyft-motion-prediction-autonomous-vehicles/discussion/193159",
  "author_name": "",
  "post_date": "2020-10-25T15:03:09.642650600Z",
  "votes": 1,
  "comment_count": 4,
  "views": 0,
  "content": "<pre><code>num_history_channels = (cfg[\"model_params\"][\"history_num_frames\"] + 1) * 2\nnum_targets = 2 * cfg[\"model_params\"][\"future_num_frames\"]\n</code></pre>\n<p>Here why we need to multiply by 2?</p>",
  "messages": [
    {
      "id": "1059880",
      "postDate": "10/25/2020 15:03:09",
      "content": "<pre><code>num_history_channels = (cfg[\"model_params\"][\"history_num_frames\"] + 1) * 2\nnum_targets = 2 * cfg[\"model_params\"][\"future_num_frames\"]\n</code></pre>\n<p>Here why we need to multiply by 2?</p>",
      "rawMarkdown": "```\nnum_history_channels = (cfg[\"model_params\"][\"history_num_frames\"] + 1) * 2\nnum_targets = 2 * cfg[\"model_params\"][\"future_num_frames\"]\n```\nHere why we need to multiply by 2?",
      "votes": null
    },
    {
      "id": "1059888",
      "postDate": "10/25/2020 15:08:19",
      "content": "<p><code>x</code> and <code>y</code> coordinates :)</p>\n<p>you need to work on making your questions harder!</p>",
      "rawMarkdown": "`x` and `y` coordinates :)\n\nyou need to work on making your questions harder!",
      "votes": null
    },
    {
      "id": "1059893",
      "postDate": "10/25/2020 15:12:30",
      "content": "<p>Ok. Thanks for the hint.  Going through the below code. It may give more clarity.<br>\n<img src=\"https://www.googleapis.com/download/storage/v1/b/kaggle-forum-message-attachments/o/inbox%2F636669%2F636bc2c289cedad13e90fe726c176406%2FScreenshot%20from%202020-10-25%2020-40-16.png?generation=1603638738285557&amp;alt=media\" alt=\"\"><br>\n\"you need to work on making your questions harder!\" - Sure.</p>",
      "rawMarkdown": "Ok. Thanks for the hint.  Going through the below code. It may give more clarity.\n![](https://www.googleapis.com/download/storage/v1/b/kaggle-forum-message-attachments/o/inbox%2F636669%2F636bc2c289cedad13e90fe726c176406%2FScreenshot%20from%202020-10-25%2020-40-16.png?generation=1603638738285557&alt=media)\n\"you need to work on making your questions harder!\" - Sure.",
      "votes": null
    },
    {
      "id": "1061672",
      "postDate": "10/27/2020 07:36:54",
      "content": "<p>Team, </p>\n<ul>\n<li>Why input to the model as  <strong>[agents_t , agents_t-1, …, agents_t-11] + [ego_agent_t, ego_agent_t-1, … ego_agent_t-11] + tl_faces_t</strong> if history_agents = 10 and include_current_frame = True ?<br>\n<img src=\"https://www.googleapis.com/download/storage/v1/b/kaggle-forum-message-attachments/o/inbox%2F636669%2F7cc4c0976d1b1c4d08703257d6419641%2FScreenshot%20from%202020-10-27%2015-27-26.png?generation=1603792711718083&amp;alt=media\" alt=\"\"></li>\n<li>Why not <strong>[add(agents_t, ego_agent_t) , add(agents_t-1,ego_agent_t-1) , …, add(agents_t-11,ego_agent_t-11)] + tl_faces_t</strong> ?</li>\n<li>How they are different for the model?</li>\n</ul>\n<p>Also, here AV treated as one of the agents in every frame.</p>",
      "rawMarkdown": "Team, \n- Why input to the model as  **[agents_t , agents_t-1, ..., agents_t-11] + [ego_agent_t, ego_agent_t-1, ... ego_agent_t-11] + tl_faces_t** if history_agents = 10 and include_current_frame = True ?\n![](https://www.googleapis.com/download/storage/v1/b/kaggle-forum-message-attachments/o/inbox%2F636669%2F7cc4c0976d1b1c4d08703257d6419641%2FScreenshot%20from%202020-10-27%2015-27-26.png?generation=1603792711718083&alt=media)\n- Why not **[add(agents_t, ego_agent_t) , add(agents_t-1,ego_agent_t-1) , ..., add(agents_t-11,ego_agent_t-11)] + tl_faces_t** ?\n- How they are different for the model?\n\nAlso, here AV treated as one of the agents in every frame.",
      "votes": null
    },
    {
      "id": "1062736",
      "postDate": "10/28/2020 06:27:07",
      "content": "<p>You may be able to drop the ego input if it is always centered in the same place, it will only give vehicle size</p>",
      "rawMarkdown": "You may be able to drop the ego input if it is always centered in the same place, it will only give vehicle size",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 1059888,
      "author_name": "zaharch",
      "author_url": "",
      "post_date": "10/25/2020 15:08:19",
      "content": "<p><code>x</code> and <code>y</code> coordinates :)</p>\n<p>you need to work on making your questions harder!</p>",
      "votes": null,
      "replies": [
        {
          "id": 1059893,
          "author_name": "pallaviroyal",
          "author_url": "",
          "post_date": "10/25/2020 15:12:30",
          "content": "<p>Ok. Thanks for the hint.  Going through the below code. It may give more clarity.<br>\n<img src=\"https://www.googleapis.com/download/storage/v1/b/kaggle-forum-message-attachments/o/inbox%2F636669%2F636bc2c289cedad13e90fe726c176406%2FScreenshot%20from%202020-10-25%2020-40-16.png?generation=1603638738285557&amp;alt=media\" alt=\"\"><br>\n\"you need to work on making your questions harder!\" - Sure.</p>",
          "votes": null,
          "replies": []
        }
      ]
    },
    {
      "id": 1061672,
      "author_name": "pallaviroyal",
      "author_url": "",
      "post_date": "10/27/2020 07:36:54",
      "content": "<p>Team, </p>\n<ul>\n<li>Why input to the model as  <strong>[agents_t , agents_t-1, …, agents_t-11] + [ego_agent_t, ego_agent_t-1, … ego_agent_t-11] + tl_faces_t</strong> if history_agents = 10 and include_current_frame = True ?<br>\n<img src=\"https://www.googleapis.com/download/storage/v1/b/kaggle-forum-message-attachments/o/inbox%2F636669%2F7cc4c0976d1b1c4d08703257d6419641%2FScreenshot%20from%202020-10-27%2015-27-26.png?generation=1603792711718083&amp;alt=media\" alt=\"\"></li>\n<li>Why not <strong>[add(agents_t, ego_agent_t) , add(agents_t-1,ego_agent_t-1) , …, add(agents_t-11,ego_agent_t-11)] + tl_faces_t</strong> ?</li>\n<li>How they are different for the model?</li>\n</ul>\n<p>Also, here AV treated as one of the agents in every frame.</p>",
      "votes": null,
      "replies": [
        {
          "id": 1062736,
          "author_name": "n3n77i",
          "author_url": "",
          "post_date": "10/28/2020 06:27:07",
          "content": "<p>You may be able to drop the ego input if it is always centered in the same place, it will only give vehicle size</p>",
          "votes": null,
          "replies": []
        }
      ]
    }
  ],
  "raw_markdown_by_id": {
    "1059880": "```\nnum_history_channels = (cfg[\"model_params\"][\"history_num_frames\"] + 1) * 2\nnum_targets = 2 * cfg[\"model_params\"][\"future_num_frames\"]\n```\nHere why we need to multiply by 2?",
    "1059888": "`x` and `y` coordinates :)\n\nyou need to work on making your questions harder!",
    "1059893": "Ok. Thanks for the hint.  Going through the below code. It may give more clarity.\n![](https://www.googleapis.com/download/storage/v1/b/kaggle-forum-message-attachments/o/inbox%2F636669%2F636bc2c289cedad13e90fe726c176406%2FScreenshot%20from%202020-10-25%2020-40-16.png?generation=1603638738285557&alt=media)\n\"you need to work on making your questions harder!\" - Sure.",
    "1061672": "Team, \n- Why input to the model as  **[agents_t , agents_t-1, ..., agents_t-11] + [ego_agent_t, ego_agent_t-1, ... ego_agent_t-11] + tl_faces_t** if history_agents = 10 and include_current_frame = True ?\n![](https://www.googleapis.com/download/storage/v1/b/kaggle-forum-message-attachments/o/inbox%2F636669%2F7cc4c0976d1b1c4d08703257d6419641%2FScreenshot%20from%202020-10-27%2015-27-26.png?generation=1603792711718083&alt=media)\n- Why not **[add(agents_t, ego_agent_t) , add(agents_t-1,ego_agent_t-1) , ..., add(agents_t-11,ego_agent_t-11)] + tl_faces_t** ?\n- How they are different for the model?\n\nAlso, here AV treated as one of the agents in every frame.",
    "1062736": "You may be able to drop the ego input if it is always centered in the same place, it will only give vehicle size"
  },
  "source": "meta"
}