{
  "id": 209048,
  "title": "How to perform Pseudo-Labeling in this Competition?",
  "url": "/competitions/cassava-leaf-disease-classification/discussion/209048",
  "author_name": "",
  "post_date": "2021-01-06T04:53:44.580605200Z",
  "votes": 1,
  "comment_count": 3,
  "views": 0,
  "content": "<p>Hi,<br>\nI learned about pseudo-labeling a couple of weeks ago and I was thinking if I would be able to apply it in this competition because the visible public test set has just one image, so it won't be possible to do any pseudo-labeling with just 1 image. Is there a way through which I could view the entire public test set? or any other way through which I could perform pseudo-labeling?<br>\nThanks.</p>",
  "messages": [
    {
      "id": "1140532",
      "postDate": "01/06/2021 04:53:44",
      "content": "<p>Hi,<br>\nI learned about pseudo-labeling a couple of weeks ago and I was thinking if I would be able to apply it in this competition because the visible public test set has just one image, so it won't be possible to do any pseudo-labeling with just 1 image. Is there a way through which I could view the entire public test set? or any other way through which I could perform pseudo-labeling?<br>\nThanks.</p>",
      "rawMarkdown": "Hi,\nI learned about pseudo-labeling a couple of weeks ago and I was thinking if I would be able to apply it in this competition because the visible public test set has just one image, so it won't be possible to do any pseudo-labeling with just 1 image. Is there a way through which I could view the entire public test set? or any other way through which I could perform pseudo-labeling?\nThanks.",
      "votes": null
    },
    {
      "id": "1140617",
      "postDate": "01/06/2021 06:06:19",
      "content": "<p>A month ago, I tried to pseudo-label the private test set, but it failed when it reached the training part, and I never tried again.</p>",
      "rawMarkdown": "A month ago, I tried to pseudo-label the private test set, but it failed when it reached the training part, and I never tried again.",
      "votes": null
    },
    {
      "id": "1141740",
      "postDate": "01/06/2021 21:33:43",
      "content": "<p>You can pseudo label the extra images from the 2019 cassava competition</p>",
      "rawMarkdown": "You can pseudo label the extra images from the 2019 cassava competition",
      "votes": null
    },
    {
      "id": "1141979",
      "postDate": "01/07/2021 03:45:55",
      "content": "<p>Ok yeah, thanks!</p>",
      "rawMarkdown": "Ok yeah, thanks!",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 1140617,
      "author_name": "zhangeng",
      "author_url": "",
      "post_date": "01/06/2021 06:06:19",
      "content": "<p>A month ago, I tried to pseudo-label the private test set, but it failed when it reached the training part, and I never tried again.</p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 1141740,
      "author_name": "yannmajewski",
      "author_url": "",
      "post_date": "01/06/2021 21:33:43",
      "content": "<p>You can pseudo label the extra images from the 2019 cassava competition</p>",
      "votes": null,
      "replies": [
        {
          "id": 1141979,
          "author_name": "aryankhatana",
          "author_url": "",
          "post_date": "01/07/2021 03:45:55",
          "content": "<p>Ok yeah, thanks!</p>",
          "votes": null,
          "replies": []
        }
      ]
    }
  ],
  "raw_markdown_by_id": {
    "1140532": "Hi,\nI learned about pseudo-labeling a couple of weeks ago and I was thinking if I would be able to apply it in this competition because the visible public test set has just one image, so it won't be possible to do any pseudo-labeling with just 1 image. Is there a way through which I could view the entire public test set? or any other way through which I could perform pseudo-labeling?\nThanks.",
    "1140617": "A month ago, I tried to pseudo-label the private test set, but it failed when it reached the training part, and I never tried again.",
    "1141740": "You can pseudo label the extra images from the 2019 cassava competition",
    "1141979": "Ok yeah, thanks!"
  },
  "source": "meta"
}