{
  "id": 105092,
  "title": "Looking for a team (LB > 1.9)",
  "url": "/competitions/champs-scalar-coupling/discussion/105092",
  "author_name": "",
  "post_date": "2019-08-21T06:00:34.776324100Z",
  "votes": 2,
  "comment_count": 4,
  "views": 0,
  "content": "<p>Hi guys, I have made some breakthroughs, but didn't have the change to submit the results yet.</p>\n\n<p>A couple of weeks ago my best single model had a LB of -1.610.\nMy 1JHN and 2JHN, for example,  had a CV of -1.19 and -2.17</p>\n\n<p>With my <strong>new model</strong>, without kfolds and fine tuning, I got -1.58 and -2.37, respectively. Great improvement.</p>\n\n<p>Gonna spend this next week fine tuning, doing 5-folds and blending the results. My model is a NN with features that I made and some tricks that gave a great boost to my score. My LGB works well too, but a little bit worse than my NN. It would be nice to join any group, but specially one with a different model, like a graph-NN.</p>\n\n<p>Thanks! :)</p>",
  "messages": [
    {
      "id": "604184",
      "postDate": "08/21/2019 06:00:34",
      "content": "<p>Hi guys, I have made some breakthroughs, but didn't have the change to submit the results yet.</p>\n\n<p>A couple of weeks ago my best single model had a LB of -1.610.\nMy 1JHN and 2JHN, for example,  had a CV of -1.19 and -2.17</p>\n\n<p>With my <strong>new model</strong>, without kfolds and fine tuning, I got -1.58 and -2.37, respectively. Great improvement.</p>\n\n<p>Gonna spend this next week fine tuning, doing 5-folds and blending the results. My model is a NN with features that I made and some tricks that gave a great boost to my score. My LGB works well too, but a little bit worse than my NN. It would be nice to join any group, but specially one with a different model, like a graph-NN.</p>\n\n<p>Thanks! :)</p>",
      "rawMarkdown": "Hi guys, I have made some breakthroughs, but didn't have the change to submit the results yet.\n\nA couple of weeks ago my best single model had a LB of -1.610.\nMy 1JHN and 2JHN, for example,  had a CV of -1.19 and -2.17\n\nWith my **new model**, without kfolds and fine tuning, I got -1.58 and -2.37, respectively. Great improvement.\n\nGonna spend this next week fine tuning, doing 5-folds and blending the results. My model is a NN with features that I made and some tricks that gave a great boost to my score. My LGB works well too, but a little bit worse than my NN. It would be nice to join any group, but specially one with a different model, like a graph-NN.\n\nThanks! :)",
      "votes": null
    },
    {
      "id": "604403",
      "postDate": "08/21/2019 10:51:37",
      "content": "<p><a href=\"/felipemello\">@felipemello</a> we have a team, let's see what team leader says <a href=\"/yassinealouini\">@yassinealouini</a></p>\n\n<p>Now we are around - 2</p>\n\n<p>Send us request and let's see.\nTeam name: MoleculesLearning</p>",
      "rawMarkdown": "felipemello we have a team, let's see what team leader says @yassinealouini\n\nNow we are around - 2\n\nSend us request and let's see.\nTeam name: MoleculesLearning",
      "votes": null
    },
    {
      "id": "604411",
      "postDate": "08/21/2019 10:56:56",
      "content": "<p>Team merge request sent <a href=\"/felipemello\">@felipemello</a>. Let me know what you think. 🙃 </p>",
      "rawMarkdown": "Team merge request sent @felipemello. Let me know what you think. 🙃",
      "votes": null
    },
    {
      "id": "604458",
      "postDate": "08/21/2019 12:16:35",
      "content": "<p>Thanks <a href=\"/yassinealouini\">@yassinealouini</a>  and <a href=\"/muhakabartay\">@muhakabartay</a>. I accepted the invite :)</p>\n\n<p>Can you guys put me in your email loop so I can share information about my model and ask about yours? My email is fmellomascarenhas@gmail.com</p>\n\n<p>Thanks!</p>",
      "rawMarkdown": "Thanks @yassinealouini  and @muhakabartay. I accepted the invite :)\n\nCan you guys put me in your email loop so I can share information about my model and ask about yours? My email is fmellomascarenhas@gmail.com\n\nThanks!",
      "votes": null
    },
    {
      "id": "604669",
      "postDate": "08/21/2019 16:27:40",
      "content": "<p>Yes. I will send you a Discord link. See you later. </p>",
      "rawMarkdown": "Yes. I will send you a Discord link. See you later.",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 604403,
      "author_name": "muhakabartay",
      "author_url": "",
      "post_date": "08/21/2019 10:51:37",
      "content": "<p><a href=\"/felipemello\">@felipemello</a> we have a team, let's see what team leader says <a href=\"/yassinealouini\">@yassinealouini</a></p>\n\n<p>Now we are around - 2</p>\n\n<p>Send us request and let's see.\nTeam name: MoleculesLearning</p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 604411,
      "author_name": "yassinealouini",
      "author_url": "",
      "post_date": "08/21/2019 10:56:56",
      "content": "<p>Team merge request sent <a href=\"/felipemello\">@felipemello</a>. Let me know what you think. 🙃 </p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 604458,
      "author_name": "felipemello",
      "author_url": "",
      "post_date": "08/21/2019 12:16:35",
      "content": "<p>Thanks <a href=\"/yassinealouini\">@yassinealouini</a>  and <a href=\"/muhakabartay\">@muhakabartay</a>. I accepted the invite :)</p>\n\n<p>Can you guys put me in your email loop so I can share information about my model and ask about yours? My email is fmellomascarenhas@gmail.com</p>\n\n<p>Thanks!</p>",
      "votes": null,
      "replies": [
        {
          "id": 604669,
          "author_name": "yassinealouini",
          "author_url": "",
          "post_date": "08/21/2019 16:27:40",
          "content": "<p>Yes. I will send you a Discord link. See you later. </p>",
          "votes": null,
          "replies": []
        }
      ]
    }
  ],
  "raw_markdown_by_id": {
    "604184": "Hi guys, I have made some breakthroughs, but didn't have the change to submit the results yet.\n\nA couple of weeks ago my best single model had a LB of -1.610.\nMy 1JHN and 2JHN, for example,  had a CV of -1.19 and -2.17\n\nWith my **new model**, without kfolds and fine tuning, I got -1.58 and -2.37, respectively. Great improvement.\n\nGonna spend this next week fine tuning, doing 5-folds and blending the results. My model is a NN with features that I made and some tricks that gave a great boost to my score. My LGB works well too, but a little bit worse than my NN. It would be nice to join any group, but specially one with a different model, like a graph-NN.\n\nThanks! :)",
    "604403": "felipemello we have a team, let's see what team leader says @yassinealouini\n\nNow we are around - 2\n\nSend us request and let's see.\nTeam name: MoleculesLearning",
    "604411": "Team merge request sent @felipemello. Let me know what you think. 🙃",
    "604458": "Thanks @yassinealouini  and @muhakabartay. I accepted the invite :)\n\nCan you guys put me in your email loop so I can share information about my model and ask about yours? My email is fmellomascarenhas@gmail.com\n\nThanks!",
    "604669": "Yes. I will send you a Discord link. See you later."
  },
  "source": "meta"
}