{
  "id": 306178,
  "title": "Homemade Yolo like model",
  "url": "/competitions/tensorflow-great-barrier-reef/discussion/306178",
  "author_name": "JeanBenoitFerry",
  "post_date": "2022-02-08T12:31:50.598000",
  "votes": 4,
  "comment_count": 5,
  "views": 0,
  "content": "<p>Hello,</p>\n<p>I tried to re-build a \"yolo like\" approach in this notebook :<br>\n<a href=\"url\" target=\"_blank\">https://www.kaggle.com/jeanbenoitferry/starfishes-model?kernelSessionId=86956961</a></p>\n<p>It is obviously very much simplified but does not seem to bring good result (LB result stays at 0).<br>\nDoes anybody have ideas on how to improve it ?</p>\n<p>Cheers !</p>",
  "messages": [
    {
      "id": 1681351,
      "postDate": "2022-02-08T12:31:50.600Z",
      "content": "<p>Hello,</p>\n<p>I tried to re-build a \"yolo like\" approach in this notebook :<br>\n<a href=\"url\" target=\"_blank\">https://www.kaggle.com/jeanbenoitferry/starfishes-model?kernelSessionId=86956961</a></p>\n<p>It is obviously very much simplified but does not seem to bring good result (LB result stays at 0).<br>\nDoes anybody have ideas on how to improve it ?</p>\n<p>Cheers !</p>",
      "rawMarkdown": "Hello,\n\nI tried to re-build a \"yolo like\" approach in this notebook :\n[https://www.kaggle.com/jeanbenoitferry/starfishes-model?kernelSessionId=86956961](url)\n\nIt is obviously very much simplified but does not seem to bring good result (LB result stays at 0).\nDoes anybody have ideas on how to improve it ?\n\nCheers !",
      "votes": 4
    },
    {
      "id": 1681595,
      "postDate": "2022-02-08T15:16:00.480Z",
      "content": "<p>I am having the same issue here. <a href=\"https://www.kaggle.com/timxinxinpeng/flexible-yolo-with-attention-asking-for-help?scriptVersionId=87129266\" target=\"_blank\">https://www.kaggle.com/timxinxinpeng/flexible-yolo-with-attention-asking-for-help?scriptVersionId=87129266</a> this model actually achieved the best performance during training among all models I trained, but only got 0 at LB. My best results on LB so far is 0.599, which means this unsuccessful submission could be beyond that. </p>",
      "rawMarkdown": "I am having the same issue here. https://www.kaggle.com/timxinxinpeng/flexible-yolo-with-attention-asking-for-help?scriptVersionId=87129266 this model actually achieved the best performance during training among all models I trained, but only got 0 at LB. My best results on LB so far is 0.599, which means this unsuccessful submission could be beyond that. ",
      "replies": [
        {
          "id": 1682193,
          "postDate": "2022-02-09T01:58:22.787Z",
          "content": "<p>did you get any score?  I see that you have many new versions.</p>",
          "rawMarkdown": "did you get any score?  I see that you have many new versions."
        },
        {
          "id": 1682230,
          "postDate": "2022-02-09T02:48:51.383Z",
          "content": "<p>Haven't yet…will keep trying, it must be the error in the submission code. </p>",
          "rawMarkdown": "Haven't yet...will keep trying, it must be the error in the submission code. "
        },
        {
          "id": 1685762,
          "postDate": "2022-02-11T14:23:42.037Z",
          "content": "<p>Finally I got to submit to LB using the modified cofig with transformer module in original YoloV5. Just a testing out, model from longer training will come out soon</p>",
          "rawMarkdown": "Finally I got to submit to LB using the modified cofig with transformer module in original YoloV5. Just a testing out, model from longer training will come out soon",
          "votes": 1
        }
      ]
    },
    {
      "id": 1682285,
      "postDate": "2022-02-09T03:46:39.527Z",
      "rawMarkdown": "",
      "isDeleted": true
    }
  ],
  "comments": [
    {
      "id": 1681595,
      "author_name": "Tim Xinxin Peng",
      "author_url": "",
      "post_date": "2022-02-08T15:16:00.480000",
      "content": "<p>I am having the same issue here. <a href=\"https://www.kaggle.com/timxinxinpeng/flexible-yolo-with-attention-asking-for-help?scriptVersionId=87129266\" target=\"_blank\">https://www.kaggle.com/timxinxinpeng/flexible-yolo-with-attention-asking-for-help?scriptVersionId=87129266</a> this model actually achieved the best performance during training among all models I trained, but only got 0 at LB. My best results on LB so far is 0.599, which means this unsuccessful submission could be beyond that. </p>",
      "votes": 0,
      "replies": [
        {
          "id": 1682193,
          "author_name": "dragon zhang",
          "author_url": "",
          "post_date": "2022-02-09T01:58:22.787000",
          "content": "<p>did you get any score?  I see that you have many new versions.</p>",
          "votes": 0,
          "replies": []
        },
        {
          "id": 1682230,
          "author_name": "Tim Xinxin Peng",
          "author_url": "",
          "post_date": "2022-02-09T02:48:51.383000",
          "content": "<p>Haven't yet…will keep trying, it must be the error in the submission code. </p>",
          "votes": 0,
          "replies": []
        },
        {
          "id": 1685762,
          "author_name": "Tim Xinxin Peng",
          "author_url": "",
          "post_date": "2022-02-11T14:23:42.037000",
          "content": "<p>Finally I got to submit to LB using the modified cofig with transformer module in original YoloV5. Just a testing out, model from longer training will come out soon</p>",
          "votes": 1,
          "replies": []
        }
      ]
    },
    {
      "id": 1682285,
      "author_name": "",
      "author_url": "",
      "post_date": "2022-02-09T03:46:39.527000",
      "content": "",
      "votes": 0,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "1681351": "Hello,\n\nI tried to re-build a \"yolo like\" approach in this notebook :\n[https://www.kaggle.com/jeanbenoitferry/starfishes-model?kernelSessionId=86956961](url)\n\nIt is obviously very much simplified but does not seem to bring good result (LB result stays at 0).\nDoes anybody have ideas on how to improve it ?\n\nCheers !",
    "1681595": "I am having the same issue here. https://www.kaggle.com/timxinxinpeng/flexible-yolo-with-attention-asking-for-help?scriptVersionId=87129266 this model actually achieved the best performance during training among all models I trained, but only got 0 at LB. My best results on LB so far is 0.599, which means this unsuccessful submission could be beyond that. ",
    "1682285": ""
  }
}