{
  "id": 197749,
  "title": "Tfrecord - Label info",
  "url": "/competitions/rfcx-species-audio-detection/discussion/197749",
  "author_name": "",
  "post_date": "2020-11-17T22:06:24.187968Z",
  "votes": 3,
  "comment_count": 1,
  "views": 0,
  "content": "<p>Hi,</p>\n<p>Firstly, thank you for proposing this challenge and for making tfrecord available.</p>\n<p>I was looking the \"label info\" in the tfrecord but I am not sure what all the informations represent. I feel there is one value which was not mention in the description.<br>\nFor instance: <br>\n\"15,1,5.4773,93.75,8.8213,1125.0,1;10,1,15.8534,947.4609,18.2973,10852.7344,0;6,1,46.4427,562.5,48.464,3187.5,0\".</p>\n<p>The first label is represented by : <br>\n15, 1, 5.4773, 93.75, 8.8213, 1125.0, 1</p>\n<p>Given    </p>\n<pre><code>tfrecords/{train,test} - competition data in the TFRecord format, which includes recording_id, audio_wav (encoded in 16-bit PCM format), and label_info (for train only), which provides a,-delimited string of the columns below (minus recording_id), where multiple labels for a recording_id are ;-delimited.\n\nColumns\n\n    recording_id - unique identifier for recording\n    species_id - unique identifier for species\n    songtype_id - unique identifier for songtype\n    t_min - start second of annotated signal\n    f_min - lower frequency of annotated signal\n    t_max - end second of annotated signal\n    f_max- upper frequency of annotated signal\n</code></pre>\n<p>So, I suppose that : <br>\n15 = species_id<br>\n1 = songtype_id<br>\n5.4773= t_min<br>\n93.75 = f_min<br>\n8.8213 = t_max<br>\n1125.0 = f_max<br>\n1 = ? </p>\n<p>So, first am I right regarding the species_id, songtype_id etc ?<br>\nWhat does the last integer represent ?</p>\n<p><a href=\"https://www.kaggle.com/ludovick/load-data-from-tfrecord\" target=\"_blank\">https://www.kaggle.com/ludovick/load-data-from-tfrecord</a> here is how I extract the label info, I am doing something wrong ?</p>",
  "messages": [
    {
      "id": "1082429",
      "postDate": "11/17/2020 22:06:24",
      "content": "<p>Hi,</p>\n<p>Firstly, thank you for proposing this challenge and for making tfrecord available.</p>\n<p>I was looking the \"label info\" in the tfrecord but I am not sure what all the informations represent. I feel there is one value which was not mention in the description.<br>\nFor instance: <br>\n\"15,1,5.4773,93.75,8.8213,1125.0,1;10,1,15.8534,947.4609,18.2973,10852.7344,0;6,1,46.4427,562.5,48.464,3187.5,0\".</p>\n<p>The first label is represented by : <br>\n15, 1, 5.4773, 93.75, 8.8213, 1125.0, 1</p>\n<p>Given    </p>\n<pre><code>tfrecords/{train,test} - competition data in the TFRecord format, which includes recording_id, audio_wav (encoded in 16-bit PCM format), and label_info (for train only), which provides a,-delimited string of the columns below (minus recording_id), where multiple labels for a recording_id are ;-delimited.\n\nColumns\n\n    recording_id - unique identifier for recording\n    species_id - unique identifier for species\n    songtype_id - unique identifier for songtype\n    t_min - start second of annotated signal\n    f_min - lower frequency of annotated signal\n    t_max - end second of annotated signal\n    f_max- upper frequency of annotated signal\n</code></pre>\n<p>So, I suppose that : <br>\n15 = species_id<br>\n1 = songtype_id<br>\n5.4773= t_min<br>\n93.75 = f_min<br>\n8.8213 = t_max<br>\n1125.0 = f_max<br>\n1 = ? </p>\n<p>So, first am I right regarding the species_id, songtype_id etc ?<br>\nWhat does the last integer represent ?</p>\n<p><a href=\"https://www.kaggle.com/ludovick/load-data-from-tfrecord\" target=\"_blank\">https://www.kaggle.com/ludovick/load-data-from-tfrecord</a> here is how I extract the label info, I am doing something wrong ?</p>",
      "rawMarkdown": "Hi,\n\nFirstly, thank you for proposing this challenge and for making tfrecord available.\n\nI was looking the \"label info\" in the tfrecord but I am not sure what all the informations represent. I feel there is one value which was not mention in the description.\nFor instance: \n\"15,1,5.4773,93.75,8.8213,1125.0,1;10,1,15.8534,947.4609,18.2973,10852.7344,0;6,1,46.4427,562.5,48.464,3187.5,0\".\n\nThe first label is represented by : \n15, 1, 5.4773, 93.75, 8.8213, 1125.0, 1\n\nGiven    \n ```\ntfrecords/{train,test} - competition data in the TFRecord format, which includes recording_id, audio_wav (encoded in 16-bit PCM format), and label_info (for train only), which provides a,-delimited string of the columns below (minus recording_id), where multiple labels for a recording_id are ;-delimited.\n\nColumns\n\n    recording_id - unique identifier for recording\n    species_id - unique identifier for species\n    songtype_id - unique identifier for songtype\n    t_min - start second of annotated signal\n    f_min - lower frequency of annotated signal\n    t_max - end second of annotated signal\n    f_max- upper frequency of annotated signal\n```\n\nSo, I suppose that : \n15 = species_id\n1 = songtype_id\n5.4773= t_min\n93.75 = f_min\n8.8213 = t_max\n1125.0 = f_max\n1 = ? \n\nSo, first am I right regarding the species_id, songtype_id etc ?\nWhat does the last integer represent ?\n\nhttps://www.kaggle.com/ludovick/load-data-from-tfrecord here is how I extract the label info, I am doing something wrong ?",
      "votes": null
    },
    {
      "id": "1082460",
      "postDate": "11/17/2020 23:13:12",
      "content": "<p>Thanks. That's an indicator of whether the label is from the <code>train_tp</code> (<code>1</code>) or <code>train_fp</code> (<code>0</code>) file.</p>\n<p>I've updated the data description.</p>",
      "rawMarkdown": "Thanks. That's an indicator of whether the label is from the `train_tp` (`1`) or `train_fp` (`0`) file.\n\nI've updated the data description.",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 1082460,
      "author_name": "inversion",
      "author_url": "",
      "post_date": "11/17/2020 23:13:12",
      "content": "<p>Thanks. That's an indicator of whether the label is from the <code>train_tp</code> (<code>1</code>) or <code>train_fp</code> (<code>0</code>) file.</p>\n<p>I've updated the data description.</p>",
      "votes": null,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "1082429": "Hi,\n\nFirstly, thank you for proposing this challenge and for making tfrecord available.\n\nI was looking the \"label info\" in the tfrecord but I am not sure what all the informations represent. I feel there is one value which was not mention in the description.\nFor instance: \n\"15,1,5.4773,93.75,8.8213,1125.0,1;10,1,15.8534,947.4609,18.2973,10852.7344,0;6,1,46.4427,562.5,48.464,3187.5,0\".\n\nThe first label is represented by : \n15, 1, 5.4773, 93.75, 8.8213, 1125.0, 1\n\nGiven    \n ```\ntfrecords/{train,test} - competition data in the TFRecord format, which includes recording_id, audio_wav (encoded in 16-bit PCM format), and label_info (for train only), which provides a,-delimited string of the columns below (minus recording_id), where multiple labels for a recording_id are ;-delimited.\n\nColumns\n\n    recording_id - unique identifier for recording\n    species_id - unique identifier for species\n    songtype_id - unique identifier for songtype\n    t_min - start second of annotated signal\n    f_min - lower frequency of annotated signal\n    t_max - end second of annotated signal\n    f_max- upper frequency of annotated signal\n```\n\nSo, I suppose that : \n15 = species_id\n1 = songtype_id\n5.4773= t_min\n93.75 = f_min\n8.8213 = t_max\n1125.0 = f_max\n1 = ? \n\nSo, first am I right regarding the species_id, songtype_id etc ?\nWhat does the last integer represent ?\n\nhttps://www.kaggle.com/ludovick/load-data-from-tfrecord here is how I extract the label info, I am doing something wrong ?",
    "1082460": "Thanks. That's an indicator of whether the label is from the `train_tp` (`1`) or `train_fp` (`0`) file.\n\nI've updated the data description."
  },
  "source": "meta"
}