{
  "id": 279365,
  "title": "Simple baseline model with LB > 0 made public",
  "url": "/competitions/sartorius-cell-instance-segmentation/discussion/279365",
  "author_name": "Julián Peller (dataista0)",
  "post_date": "2021-10-18T01:10:09.587000",
  "votes": 17,
  "comment_count": 0,
  "views": 0,
  "content": "<p>Hello community,<br>\nI just made public a notebook with LB &gt; 0 (though very modest). You can find it <a href=\"https://www.kaggle.com/julian3833/sartorius-starter-baseline-torch-u-net\" target=\"_blank\">here</a>.<br>\nIt is very simple and very basic (it doesn't even have a validation set lol), but it covers the pipeline fully.<br>\nI saw that there are a lot of 0.0 submissions, so it might help to debug some problems in the pipelines. I have struggled quite a bit with the run-length encoding and decoding, but I think I have it going. It also takes care of separating the predictions into different components. It's done roughly, but it's done.<br>\nHopefully, some of you might find some parts of the notebook useful.<br>\nHappy kaggling!<br>\ndataista0</p>",
  "messages": [
    {
      "id": 1548094,
      "postDate": "2021-10-18T01:10:09.587Z",
      "content": "<p>Hello community,<br>\nI just made public a notebook with LB &gt; 0 (though very modest). You can find it <a href=\"https://www.kaggle.com/julian3833/sartorius-starter-baseline-torch-u-net\" target=\"_blank\">here</a>.<br>\nIt is very simple and very basic (it doesn't even have a validation set lol), but it covers the pipeline fully.<br>\nI saw that there are a lot of 0.0 submissions, so it might help to debug some problems in the pipelines. I have struggled quite a bit with the run-length encoding and decoding, but I think I have it going. It also takes care of separating the predictions into different components. It's done roughly, but it's done.<br>\nHopefully, some of you might find some parts of the notebook useful.<br>\nHappy kaggling!<br>\ndataista0</p>",
      "rawMarkdown": "\nHello community,\nI just made public a notebook with LB > 0 (though very modest). You can find it [here](https://www.kaggle.com/julian3833/sartorius-starter-baseline-torch-u-net).\n\n\nIt is very simple and very basic (it doesn't even have a validation set lol), but it covers the pipeline fully.\n\n\nI saw that there are a lot of 0.0 submissions, so it might help to debug some problems in the pipelines. I have struggled quite a bit with the run-length encoding and decoding, but I think I have it going. It also takes care of separating the predictions into different components. It's done roughly, but it's done.\n\n\nHopefully, some of you might find some parts of the notebook useful.\n\n\nHappy kaggling!\n\ndataista0\n",
      "votes": 17
    }
  ],
  "comments": [],
  "raw_markdown_by_id": {
    "1548094": "\nHello community,\nI just made public a notebook with LB > 0 (though very modest). You can find it [here](https://www.kaggle.com/julian3833/sartorius-starter-baseline-torch-u-net).\n\n\nIt is very simple and very basic (it doesn't even have a validation set lol), but it covers the pipeline fully.\n\n\nI saw that there are a lot of 0.0 submissions, so it might help to debug some problems in the pipelines. I have struggled quite a bit with the run-length encoding and decoding, but I think I have it going. It also takes care of separating the predictions into different components. It's done roughly, but it's done.\n\n\nHopefully, some of you might find some parts of the notebook useful.\n\n\nHappy kaggling!\n\ndataista0\n"
  }
}