{
  "id": 412528,
  "title": "Updated reference code with real submission ",
  "url": "/competitions/google-research-identify-contrails-reduce-global-warming/discussion/412528",
  "author_name": "Nirjhar Roy",
  "post_date": "2023-05-24T05:15:14.831000",
  "votes": 5,
  "comment_count": 3,
  "views": 0,
  "content": "<p>Hello - after couple of failed submissions , I have taken some reference from the existing notebooks and created a real submission . please refer and use if you are beginner and stuck with submission failures . </p>\n<p>Notebook path :<br>\n<a href=\"https://www.kaggle.com/code/phoenix9032/inference-unet-effnetb0-on-ash-v2-baseline\" target=\"_blank\">https://www.kaggle.com/code/phoenix9032/inference-unet-effnetb0-on-ash-v2-baseline</a></p>",
  "messages": [
    {
      "id": 2271683,
      "postDate": "2023-05-24T05:15:14.830Z",
      "content": "<p>Hello - after couple of failed submissions , I have taken some reference from the existing notebooks and created a real submission . please refer and use if you are beginner and stuck with submission failures . </p>\n<p>Notebook path :<br>\n<a href=\"https://www.kaggle.com/code/phoenix9032/inference-unet-effnetb0-on-ash-v2-baseline\" target=\"_blank\">https://www.kaggle.com/code/phoenix9032/inference-unet-effnetb0-on-ash-v2-baseline</a></p>",
      "rawMarkdown": "Hello - after couple of failed submissions , I have taken some reference from the existing notebooks and created a real submission . please refer and use if you are beginner and stuck with submission failures . \n\nNotebook path :\nhttps://www.kaggle.com/code/phoenix9032/inference-unet-effnetb0-on-ash-v2-baseline",
      "votes": 3
    },
    {
      "id": 2271737,
      "postDate": "2023-05-24T05:56:30.783Z",
      "content": "<p>Hi, I've successfully submitted several notebooks, but the lb score is extremely lower than my cv score eg.0.4 -&gt; 0.017<br>\nAre there any wrong in a submission process like this: image tensor[1, 1, 256, 256] -squeeze-&gt; image tensor[256, 256] --&gt; image array[256, 256] -rle_encode&amp;list2str-&gt; a string to submit<br>\nThank u!</p>",
      "rawMarkdown": "Hi, I've successfully submitted several notebooks, but the lb score is extremely lower than my cv score eg.0.4 -> 0.017\nAre there any wrong in a submission process like this: image tensor[1, 1, 256, 256] -squeeze-> image tensor[256, 256] --> image array[256, 256] -rle_encode&list2str-> a string to submit\nThank u!",
      "votes": 1,
      "replies": [
        {
          "id": 2271977,
          "postDate": "2023-05-24T08:14:19.750Z",
          "content": "<p>Hello - Here is what I did to debug my 0.000 score . <br>\nRun the predictions on validation dataset to generate rle for validation data . <br>\nConvert back the generated rle back to mask using rle_decode and plot those masks on the validation data and compare results with the validation given masks . </p>",
          "rawMarkdown": "Hello - Here is what I did to debug my 0.000 score . \nRun the predictions on validation dataset to generate rle for validation data . \nConvert back the generated rle back to mask using rle_decode and plot those masks on the validation data and compare results with the validation given masks . ",
          "votes": 1
        },
        {
          "id": 2282904,
          "postDate": "2023-05-31T22:26:50.913Z",
          "content": "<p><a href=\"https://www.kaggle.com/lupin11\" target=\"_blank\">@lupin11</a> what did the problem end up being? I think I'm having a similar issue.</p>",
          "rawMarkdown": "@lupin11 what did the problem end up being? I think I'm having a similar issue."
        }
      ]
    }
  ],
  "comments": [
    {
      "id": 2271737,
      "author_name": "LUPIN11",
      "author_url": "",
      "post_date": "2023-05-24T05:56:30.783000",
      "content": "<p>Hi, I've successfully submitted several notebooks, but the lb score is extremely lower than my cv score eg.0.4 -&gt; 0.017<br>\nAre there any wrong in a submission process like this: image tensor[1, 1, 256, 256] -squeeze-&gt; image tensor[256, 256] --&gt; image array[256, 256] -rle_encode&amp;list2str-&gt; a string to submit<br>\nThank u!</p>",
      "votes": 1,
      "replies": [
        {
          "id": 2271977,
          "author_name": "Nirjhar Roy",
          "author_url": "",
          "post_date": "2023-05-24T08:14:19.750000",
          "content": "<p>Hello - Here is what I did to debug my 0.000 score . <br>\nRun the predictions on validation dataset to generate rle for validation data . <br>\nConvert back the generated rle back to mask using rle_decode and plot those masks on the validation data and compare results with the validation given masks . </p>",
          "votes": 1,
          "replies": []
        },
        {
          "id": 2282904,
          "author_name": "Yousef Rabi",
          "author_url": "",
          "post_date": "2023-05-31T22:26:50.913000",
          "content": "<p><a href=\"https://www.kaggle.com/lupin11\" target=\"_blank\">@lupin11</a> what did the problem end up being? I think I'm having a similar issue.</p>",
          "votes": 0,
          "replies": []
        }
      ]
    }
  ],
  "raw_markdown_by_id": {
    "2271683": "Hello - after couple of failed submissions , I have taken some reference from the existing notebooks and created a real submission . please refer and use if you are beginner and stuck with submission failures . \n\nNotebook path :\nhttps://www.kaggle.com/code/phoenix9032/inference-unet-effnetb0-on-ash-v2-baseline",
    "2271737": "Hi, I've successfully submitted several notebooks, but the lb score is extremely lower than my cv score eg.0.4 -> 0.017\nAre there any wrong in a submission process like this: image tensor[1, 1, 256, 256] -squeeze-> image tensor[256, 256] --> image array[256, 256] -rle_encode&list2str-> a string to submit\nThank u!"
  }
}