{
  "id": 296257,
  "title": "Is there a way to continue training in YOLOx?",
  "url": "/competitions/tensorflow-great-barrier-reef/discussion/296257",
  "author_name": "",
  "post_date": "2021-12-20T17:15:23.387612300Z",
  "votes": null,
  "comment_count": 2,
  "views": 0,
  "content": "<p>I am new to machine learning and YOLOx. I know there is a method to continue training \"detectron2\".<br>\nI couldn't find anything similar in YOLOx. <br>\nLet's say I trained my model for 30 epoch due to 9 hour limit, if I use best model.pth from that training and use it in next training session for 30 more epoch will it be same as training for 60 epoch? Sorry if it sounds silly question but I really have no idea. <br>\nAny help is appreciated.</p>",
  "messages": [
    {
      "id": "1624254",
      "postDate": "12/20/2021 17:15:23",
      "content": "<p>I am new to machine learning and YOLOx. I know there is a method to continue training \"detectron2\".<br>\nI couldn't find anything similar in YOLOx. <br>\nLet's say I trained my model for 30 epoch due to 9 hour limit, if I use best model.pth from that training and use it in next training session for 30 more epoch will it be same as training for 60 epoch? Sorry if it sounds silly question but I really have no idea. <br>\nAny help is appreciated.</p>",
      "rawMarkdown": "I am new to machine learning and YOLOx. I know there is a method to continue training \"detectron2\".\nI couldn't find anything similar in YOLOx. \nLet's say I trained my model for 30 epoch due to 9 hour limit, if I use best model.pth from that training and use it in next training session for 30 more epoch will it be same as training for 60 epoch? Sorry if it sounds silly question but I really have no idea. \nAny help is appreciated.",
      "votes": null
    },
    {
      "id": "1624442",
      "postDate": "12/20/2021 22:27:21",
      "content": "<p>Yes. You can use --resume option.</p>",
      "rawMarkdown": "Yes. You can use --resume option.",
      "votes": null
    },
    {
      "id": "1625045",
      "postDate": "12/21/2021 13:23:58",
      "content": "<p>Thank you so much.</p>",
      "rawMarkdown": "Thank you so much.",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 1624442,
      "author_name": "remekkinas",
      "author_url": "",
      "post_date": "12/20/2021 22:27:21",
      "content": "<p>Yes. You can use --resume option.</p>",
      "votes": null,
      "replies": [
        {
          "id": 1625045,
          "author_name": "keshhere",
          "author_url": "",
          "post_date": "12/21/2021 13:23:58",
          "content": "<p>Thank you so much.</p>",
          "votes": null,
          "replies": []
        }
      ]
    }
  ],
  "raw_markdown_by_id": {
    "1624254": "I am new to machine learning and YOLOx. I know there is a method to continue training \"detectron2\".\nI couldn't find anything similar in YOLOx. \nLet's say I trained my model for 30 epoch due to 9 hour limit, if I use best model.pth from that training and use it in next training session for 30 more epoch will it be same as training for 60 epoch? Sorry if it sounds silly question but I really have no idea. \nAny help is appreciated.",
    "1624442": "Yes. You can use --resume option.",
    "1625045": "Thank you so much."
  },
  "source": "meta"
}