{
  "id": 241280,
  "title": "efficientnetv2 is out. anyone would try it?",
  "url": "/competitions/siim-covid19-detection/discussion/241280",
  "author_name": "",
  "post_date": "2021-05-23T22:51:22.452992600Z",
  "votes": 3,
  "comment_count": 9,
  "views": 0,
  "content": "<p>since V2 is much faster than V1.</p>",
  "messages": [
    {
      "id": "1320242",
      "postDate": "05/23/2021 22:51:22",
      "content": "<p>since V2 is much faster than V1.</p>",
      "rawMarkdown": "since V2 is much faster than V1.",
      "votes": null
    },
    {
      "id": "1320339",
      "postDate": "05/24/2021 02:42:31",
      "content": "<p>Yes. For now, there was not much difference in both score and speed in the data of this competition.<br>\nv1 b0: 1epoch 41sec<br>\nv2 b0: 1epoch 41.4sec</p>",
      "rawMarkdown": "Yes. For now, there was not much difference in both score and speed in the data of this competition.\nv1 b0: 1epoch 41sec\nv2 b0: 1epoch 41.4sec",
      "votes": null
    },
    {
      "id": "1320528",
      "postDate": "05/24/2021 06:30:44",
      "content": "<p>Could you share the starter? I would like to try. Thanks</p>",
      "rawMarkdown": "Could you share the starter? I would like to try. Thanks",
      "votes": null
    },
    {
      "id": "1320578",
      "postDate": "05/24/2021 07:08:38",
      "content": "<p>You can easily try it with the latest version of timm. It is a good idea to combine it with the starter that is out now.</p>",
      "rawMarkdown": "You can easily try it with the latest version of timm. It is a good idea to combine it with the starter that is out now.",
      "votes": null
    },
    {
      "id": "1320595",
      "postDate": "05/24/2021 07:27:17",
      "content": "<p>I would be sharing a starter notebook shortly if I have time. </p>",
      "rawMarkdown": "I would be sharing a starter notebook shortly if I have time.",
      "votes": null
    },
    {
      "id": "1320676",
      "postDate": "05/24/2021 08:38:32",
      "content": "<p>Thanks for your share.  I don't know that It is out already!</p>",
      "rawMarkdown": "Thanks for your share.  I don't know that It is out already!",
      "votes": null
    },
    {
      "id": "1320694",
      "postDate": "05/24/2021 08:49:44",
      "content": "<p>timms latest version includes efficientNetv2</p>",
      "rawMarkdown": "timms latest version includes efficientNetv2",
      "votes": null
    },
    {
      "id": "1321656",
      "postDate": "05/24/2021 20:37:30",
      "content": "<p>I'll post a Tensorflow starter shortly. Keep an eye out.</p>",
      "rawMarkdown": "I'll post a Tensorflow starter shortly. Keep an eye out.",
      "votes": null
    },
    {
      "id": "1321657",
      "postDate": "05/24/2021 20:38:19",
      "content": "<p>It shocks me to see no speed improvement. Unless it's being bottlenecked by the I/O?</p>",
      "rawMarkdown": "It shocks me to see no speed improvement. Unless it's being bottlenecked by the I/O?",
      "votes": null
    },
    {
      "id": "1321668",
      "postDate": "05/24/2021 20:50:18",
      "content": "<p>I use efficientnetv2  in SETI  without pretrained weights, training from scratch, result is not attractive.  When I use pretrained weight,  It reports some error, probably due to its different operation.  Not attempt to crack the code yet.</p>",
      "rawMarkdown": "I use efficientnetv2  in SETI  without pretrained weights, training from scratch, result is not attractive.  When I use pretrained weight,  It reports some error, probably due to its different operation.  Not attempt to crack the code yet.",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 1320339,
      "author_name": "yujiariyasu",
      "author_url": "",
      "post_date": "05/24/2021 02:42:31",
      "content": "<p>Yes. For now, there was not much difference in both score and speed in the data of this competition.<br>\nv1 b0: 1epoch 41sec<br>\nv2 b0: 1epoch 41.4sec</p>",
      "votes": null,
      "replies": [
        {
          "id": 1320528,
          "author_name": "dragonzhang",
          "author_url": "",
          "post_date": "05/24/2021 06:30:44",
          "content": "<p>Could you share the starter? I would like to try. Thanks</p>",
          "votes": null,
          "replies": []
        },
        {
          "id": 1320578,
          "author_name": "yujiariyasu",
          "author_url": "",
          "post_date": "05/24/2021 07:08:38",
          "content": "<p>You can easily try it with the latest version of timm. It is a good idea to combine it with the starter that is out now.</p>",
          "votes": null,
          "replies": []
        },
        {
          "id": 1320595,
          "author_name": "morizin",
          "author_url": "",
          "post_date": "05/24/2021 07:27:17",
          "content": "<p>I would be sharing a starter notebook shortly if I have time. </p>",
          "votes": null,
          "replies": []
        },
        {
          "id": 1320676,
          "author_name": "dragonzhang",
          "author_url": "",
          "post_date": "05/24/2021 08:38:32",
          "content": "<p>Thanks for your share.  I don't know that It is out already!</p>",
          "votes": null,
          "replies": []
        },
        {
          "id": 1320694,
          "author_name": "morizin",
          "author_url": "",
          "post_date": "05/24/2021 08:49:44",
          "content": "<p>timms latest version includes efficientNetv2</p>",
          "votes": null,
          "replies": []
        },
        {
          "id": 1321657,
          "author_name": "dschettler8845",
          "author_url": "",
          "post_date": "05/24/2021 20:38:19",
          "content": "<p>It shocks me to see no speed improvement. Unless it's being bottlenecked by the I/O?</p>",
          "votes": null,
          "replies": []
        },
        {
          "id": 1321668,
          "author_name": "dragonzhang",
          "author_url": "",
          "post_date": "05/24/2021 20:50:18",
          "content": "<p>I use efficientnetv2  in SETI  without pretrained weights, training from scratch, result is not attractive.  When I use pretrained weight,  It reports some error, probably due to its different operation.  Not attempt to crack the code yet.</p>",
          "votes": null,
          "replies": []
        }
      ]
    },
    {
      "id": 1321656,
      "author_name": "dschettler8845",
      "author_url": "",
      "post_date": "05/24/2021 20:37:30",
      "content": "<p>I'll post a Tensorflow starter shortly. Keep an eye out.</p>",
      "votes": null,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "1320242": "since V2 is much faster than V1.",
    "1320339": "Yes. For now, there was not much difference in both score and speed in the data of this competition.\nv1 b0: 1epoch 41sec\nv2 b0: 1epoch 41.4sec",
    "1320528": "Could you share the starter? I would like to try. Thanks",
    "1320578": "You can easily try it with the latest version of timm. It is a good idea to combine it with the starter that is out now.",
    "1320595": "I would be sharing a starter notebook shortly if I have time.",
    "1320676": "Thanks for your share.  I don't know that It is out already!",
    "1320694": "timms latest version includes efficientNetv2",
    "1321656": "I'll post a Tensorflow starter shortly. Keep an eye out.",
    "1321657": "It shocks me to see no speed improvement. Unless it's being bottlenecked by the I/O?",
    "1321668": "I use efficientnetv2  in SETI  without pretrained weights, training from scratch, result is not attractive.  When I use pretrained weight,  It reports some error, probably due to its different operation.  Not attempt to crack the code yet."
  },
  "source": "meta"
}