{
  "id": 188025,
  "title": "Submission CSV Not Found ",
  "url": "/competitions/global-wheat-detection/discussion/188025",
  "author_name": "",
  "post_date": "2020-10-01T10:19:25.368932300Z",
  "votes": null,
  "comment_count": 2,
  "views": 0,
  "content": "<p>Hi!</p>\n<p>I have been working all week trying to get a simple YOLOv5 submission working and would really appreciate any advise on how to avoid the 'Submission CSV Not Found' error.</p>\n<p>My code is working completely fine on the small test set they give us, this is how I am running detect.py (the only place I reference the test set):</p>\n<p>%run detect.py --weights weights/best.pt --img 1024 --conf 0.4 --source ../input/global-wheat-detection/test --save-txt --device '0'</p>\n<p>This is where I am writing my submission csv (which has the correct prediction string given the small test set:</p>\n<p>submission_dir = '/kaggle/working'<br>\nresultFrame.to_csv(submission_dir + \"/submission.csv\", index = False)</p>\n<p>And I have attached my dataframe so you can see there are no issues.</p>\n<p>I have also made sure to include empty prediction strings for the image_id's that get no predictions.</p>\n<p>I would really really appreciate some tips as this has been extremely demotivating for me!</p>\n<p>Thanks!</p>",
  "messages": [
    {
      "id": "1033863",
      "postDate": "10/01/2020 10:19:25",
      "content": "<p>Hi!</p>\n<p>I have been working all week trying to get a simple YOLOv5 submission working and would really appreciate any advise on how to avoid the 'Submission CSV Not Found' error.</p>\n<p>My code is working completely fine on the small test set they give us, this is how I am running detect.py (the only place I reference the test set):</p>\n<p>%run detect.py --weights weights/best.pt --img 1024 --conf 0.4 --source ../input/global-wheat-detection/test --save-txt --device '0'</p>\n<p>This is where I am writing my submission csv (which has the correct prediction string given the small test set:</p>\n<p>submission_dir = '/kaggle/working'<br>\nresultFrame.to_csv(submission_dir + \"/submission.csv\", index = False)</p>\n<p>And I have attached my dataframe so you can see there are no issues.</p>\n<p>I have also made sure to include empty prediction strings for the image_id's that get no predictions.</p>\n<p>I would really really appreciate some tips as this has been extremely demotivating for me!</p>\n<p>Thanks!</p>",
      "rawMarkdown": "Hi!\n\nI have been working all week trying to get a simple YOLOv5 submission working and would really appreciate any advise on how to avoid the 'Submission CSV Not Found' error.\n\nMy code is working completely fine on the small test set they give us, this is how I am running detect.py (the only place I reference the test set):\n\n%run detect.py --weights weights/best.pt --img 1024 --conf 0.4 --source ../input/global-wheat-detection/test --save-txt --device '0'\n\nThis is where I am writing my submission csv (which has the correct prediction string given the small test set:\n\nsubmission_dir = '/kaggle/working'\nresultFrame.to_csv(submission_dir + \"/submission.csv\", index = False)\n\n\nAnd I have attached my dataframe so you can see there are no issues.\n\nI have also made sure to include empty prediction strings for the image_id's that get no predictions.\n\nI would really really appreciate some tips as this has been extremely demotivating for me!\n\nThanks!",
      "votes": null
    },
    {
      "id": "1033876",
      "postDate": "10/01/2020 10:30:02",
      "content": "<p>If it helps, here is my notebook <a href=\"https://www.kaggle.com/kaelanlockhart/yolov5works\" target=\"_blank\">https://www.kaggle.com/kaelanlockhart/yolov5works</a></p>",
      "rawMarkdown": "If it helps, here is my notebook https://www.kaggle.com/kaelanlockhart/yolov5works",
      "votes": null
    },
    {
      "id": "1034594",
      "postDate": "10/02/2020 00:15:16",
      "content": "<p>Sorry, should have searched the other posts first, for anyone that comes after me, just resize the test images! See my notebook below for the code</p>",
      "rawMarkdown": "Sorry, should have searched the other posts first, for anyone that comes after me, just resize the test images! See my notebook below for the code",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 1033876,
      "author_name": "kaelanlockhart",
      "author_url": "",
      "post_date": "10/01/2020 10:30:02",
      "content": "<p>If it helps, here is my notebook <a href=\"https://www.kaggle.com/kaelanlockhart/yolov5works\" target=\"_blank\">https://www.kaggle.com/kaelanlockhart/yolov5works</a></p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 1034594,
      "author_name": "kaelanlockhart",
      "author_url": "",
      "post_date": "10/02/2020 00:15:16",
      "content": "<p>Sorry, should have searched the other posts first, for anyone that comes after me, just resize the test images! See my notebook below for the code</p>",
      "votes": null,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "1033863": "Hi!\n\nI have been working all week trying to get a simple YOLOv5 submission working and would really appreciate any advise on how to avoid the 'Submission CSV Not Found' error.\n\nMy code is working completely fine on the small test set they give us, this is how I am running detect.py (the only place I reference the test set):\n\n%run detect.py --weights weights/best.pt --img 1024 --conf 0.4 --source ../input/global-wheat-detection/test --save-txt --device '0'\n\nThis is where I am writing my submission csv (which has the correct prediction string given the small test set:\n\nsubmission_dir = '/kaggle/working'\nresultFrame.to_csv(submission_dir + \"/submission.csv\", index = False)\n\n\nAnd I have attached my dataframe so you can see there are no issues.\n\nI have also made sure to include empty prediction strings for the image_id's that get no predictions.\n\nI would really really appreciate some tips as this has been extremely demotivating for me!\n\nThanks!",
    "1033876": "If it helps, here is my notebook https://www.kaggle.com/kaelanlockhart/yolov5works",
    "1034594": "Sorry, should have searched the other posts first, for anyone that comes after me, just resize the test images! See my notebook below for the code"
  },
  "source": "meta"
}