{
  "id": 166793,
  "title": "Leaderboard of Compliant Submissions",
  "url": "/competitions/global-wheat-detection/discussion/166793",
  "author_name": "Matt",
  "post_date": "2020-07-14T03:22:54.937000",
  "votes": 84,
  "comment_count": 122,
  "views": 0,
  "content": "<p>Right now it's very difficult to tell what the current leaderboard standings are due to <a href=\"https://www.kaggle.com/c/global-wheat-detection/discussion/165735\">the licensing issue</a>. As <a href=\"/alexandersoare\">@alexandersoare</a> suggested (<a href=\"https://www.kaggle.com/c/global-wheat-detection/discussion/165424\">thread</a>), some competitors have changed their team name to indicate their compliance. However, it's still hard to tell as some people have compliant solutions but their highest LB score isn't. So, I would like to use this thread to gather the highest compliant LB scores. The current standings (to the best of my knowledge) are as follows:</p>\n\n<ol>\n<li>0.7753 <a href=\"/tereka\">@tereka</a> </li>\n<li>0.7736 <a href=\"/amiras\">@amiras</a> </li>\n<li>0.7733 <a href=\"/alexanderliao\">@alexanderliao</a> </li>\n<li>0.7695 <a href=\"/datahobbit\">@datahobbit</a> and team</li>\n<li>0.7690 <a href=\"/nguyenbadung\">@nguyenbadung</a> </li>\n<li>0.7680 <a href=\"/chanhu\">@chanhu</a> </li>\n<li>0.7662 <a href=\"/ratthachat\">@ratthachat</a> </li>\n<li>0.7630 <a href=\"/nvnnghia\">@nvnnghia</a> and team</li>\n<li>0.7625 <a href=\"/li325040229\">@li325040229</a> </li>\n<li>0.7602 <a href=\"/watzisname\">@watzisname</a> </li>\n<li>0.7601 <a href=\"/wasupandceacar\">@wasupandceacar</a> </li>\n<li>0.7601 <a href=\"/raufyagfarov\">@raufyagfarov</a> </li>\n<li>0.7597 <a href=\"/ciel42\">@ciel42</a> </li>\n<li>0.7591 <a href=\"/creatrol\">@creatrol</a> </li>\n<li>0.7582 <a href=\"/guizengyou\">@guizengyou</a> </li>\n<li>0.7560 <a href=\"/nacirbouazizi\">@nacirbouazizi</a> </li>\n<li>0.7553 <a href=\"/phalanx\">@phalanx</a> </li>\n<li>0.7552 <a href=\"/alanchen1115\">@alanchen1115</a> </li>\n<li>0.7550 <a href=\"/voglinio\">@voglinio</a></li>\n<li>0.7548 <a href=\"/marcinczelej\">@marcinczelej</a> </li>\n<li>0.7541 <a href=\"/bendang\">@bendang</a> </li>\n<li>0.7536 <a href=\"/alexandersoare\">@alexandersoare</a> </li>\n<li>0.7529 <a href=\"/reppic\">@reppic</a> </li>\n<li>0.7514 <a href=\"/vishnus\">@vishnus</a> </li>\n<li>0.7509 <a href=\"/tiagotml\">@tiagotml</a> </li>\n<li>0.7496 <a href=\"/stanleyjzheng\">@stanleyjzheng</a> </li>\n<li>0.7491 <a href=\"/sanantoha\">@sanantoha</a> </li>\n<li>0.7484 <a href=\"/jihangz\">@jihangz</a> </li>\n<li>0.7479 <a href=\"/whurobin\">@whurobin</a> </li>\n<li>0.7464 myself</li>\n<li>0.7451 <a href=\"/ufownl\">@ufownl</a> </li>\n<li>0.7450 <a href=\"/gaolijing\">@gaolijing</a> </li>\n<li>0.7438 <a href=\"/truonghoang\">@truonghoang</a> </li>\n<li>0.7434 <a href=\"/zhaonat\">@zhaonat</a> </li>\n<li>0.7428 <a href=\"/alanchen1115\">@alanchen1115</a> and team</li>\n<li>0.7398 <a href=\"/cswwp347724\">@cswwp347724</a> </li>\n<li>0.7384 <a href=\"/manojprabhaakr\">@manojprabhaakr</a> </li>\n<li>0.7382 <a href=\"/thomasdic2000\">@thomasdic2000</a> </li>\n<li>0.7379 <a href=\"/sistmrl\">@sistmrl</a> </li>\n</ol>\n\n<p>This is obviously an incomplete list and I have undoubtedly missed many of you. If you'd like to share your highest compliant LB score, just comment below and I will try to keep the list up to date!</p>",
  "messages": [
    {
      "id": 928485,
      "postDate": "2020-07-14T03:22:54.937Z",
      "content": "<p>Right now it's very difficult to tell what the current leaderboard standings are due to <a href=\"https://www.kaggle.com/c/global-wheat-detection/discussion/165735\">the licensing issue</a>. As <a href=\"/alexandersoare\">@alexandersoare</a> suggested (<a href=\"https://www.kaggle.com/c/global-wheat-detection/discussion/165424\">thread</a>), some competitors have changed their team name to indicate their compliance. However, it's still hard to tell as some people have compliant solutions but their highest LB score isn't. So, I would like to use this thread to gather the highest compliant LB scores. The current standings (to the best of my knowledge) are as follows:</p>\n\n<ol>\n<li>0.7753 <a href=\"/tereka\">@tereka</a> </li>\n<li>0.7736 <a href=\"/amiras\">@amiras</a> </li>\n<li>0.7733 <a href=\"/alexanderliao\">@alexanderliao</a> </li>\n<li>0.7695 <a href=\"/datahobbit\">@datahobbit</a> and team</li>\n<li>0.7690 <a href=\"/nguyenbadung\">@nguyenbadung</a> </li>\n<li>0.7680 <a href=\"/chanhu\">@chanhu</a> </li>\n<li>0.7662 <a href=\"/ratthachat\">@ratthachat</a> </li>\n<li>0.7630 <a href=\"/nvnnghia\">@nvnnghia</a> and team</li>\n<li>0.7625 <a href=\"/li325040229\">@li325040229</a> </li>\n<li>0.7602 <a href=\"/watzisname\">@watzisname</a> </li>\n<li>0.7601 <a href=\"/wasupandceacar\">@wasupandceacar</a> </li>\n<li>0.7601 <a href=\"/raufyagfarov\">@raufyagfarov</a> </li>\n<li>0.7597 <a href=\"/ciel42\">@ciel42</a> </li>\n<li>0.7591 <a href=\"/creatrol\">@creatrol</a> </li>\n<li>0.7582 <a href=\"/guizengyou\">@guizengyou</a> </li>\n<li>0.7560 <a href=\"/nacirbouazizi\">@nacirbouazizi</a> </li>\n<li>0.7553 <a href=\"/phalanx\">@phalanx</a> </li>\n<li>0.7552 <a href=\"/alanchen1115\">@alanchen1115</a> </li>\n<li>0.7550 <a href=\"/voglinio\">@voglinio</a></li>\n<li>0.7548 <a href=\"/marcinczelej\">@marcinczelej</a> </li>\n<li>0.7541 <a href=\"/bendang\">@bendang</a> </li>\n<li>0.7536 <a href=\"/alexandersoare\">@alexandersoare</a> </li>\n<li>0.7529 <a href=\"/reppic\">@reppic</a> </li>\n<li>0.7514 <a href=\"/vishnus\">@vishnus</a> </li>\n<li>0.7509 <a href=\"/tiagotml\">@tiagotml</a> </li>\n<li>0.7496 <a href=\"/stanleyjzheng\">@stanleyjzheng</a> </li>\n<li>0.7491 <a href=\"/sanantoha\">@sanantoha</a> </li>\n<li>0.7484 <a href=\"/jihangz\">@jihangz</a> </li>\n<li>0.7479 <a href=\"/whurobin\">@whurobin</a> </li>\n<li>0.7464 myself</li>\n<li>0.7451 <a href=\"/ufownl\">@ufownl</a> </li>\n<li>0.7450 <a href=\"/gaolijing\">@gaolijing</a> </li>\n<li>0.7438 <a href=\"/truonghoang\">@truonghoang</a> </li>\n<li>0.7434 <a href=\"/zhaonat\">@zhaonat</a> </li>\n<li>0.7428 <a href=\"/alanchen1115\">@alanchen1115</a> and team</li>\n<li>0.7398 <a href=\"/cswwp347724\">@cswwp347724</a> </li>\n<li>0.7384 <a href=\"/manojprabhaakr\">@manojprabhaakr</a> </li>\n<li>0.7382 <a href=\"/thomasdic2000\">@thomasdic2000</a> </li>\n<li>0.7379 <a href=\"/sistmrl\">@sistmrl</a> </li>\n</ol>\n\n<p>This is obviously an incomplete list and I have undoubtedly missed many of you. If you'd like to share your highest compliant LB score, just comment below and I will try to keep the list up to date!</p>",
      "rawMarkdown": "Right now it's very difficult to tell what the current leaderboard standings are due to [the licensing issue](https://www.kaggle.com/c/global-wheat-detection/discussion/165735). As @alexandersoare suggested ([thread](https://www.kaggle.com/c/global-wheat-detection/discussion/165424)), some competitors have changed their team name to indicate their compliance. However, it's still hard to tell as some people have compliant solutions but their highest LB score isn't. So, I would like to use this thread to gather the highest compliant LB scores. The current standings (to the best of my knowledge) are as follows:\n\n1. 0.7753 @tereka \n2. 0.7736 @amiras \n3. 0.7733 @alexanderliao \n4. 0.7695 @datahobbit and team\n5. 0.7690 @nguyenbadung \n6. 0.7680 @chanhu \n7. 0.7662 @ratthachat \n8. 0.7630 @nvnnghia and team\n9. 0.7625 @li325040229 \n10. 0.7602 @watzisname \n11. 0.7601 @wasupandceacar \n12. 0.7601 @raufyagfarov \n13. 0.7597 @ciel42 \n14. 0.7591 @creatrol \n15. 0.7582 @guizengyou \n16. 0.7560 @nacirbouazizi \n17. 0.7553 @phalanx \n18. 0.7552 @alanchen1115 \n19. 0.7550 @voglinio\n20. 0.7548 @marcinczelej \n21. 0.7541 @bendang \n22. 0.7536 @alexandersoare \n23. 0.7529 @reppic \n24. 0.7514 @vishnus \n25. 0.7509 @tiagotml \n26. 0.7496 @stanleyjzheng \n27. 0.7491 @sanantoha \n28. 0.7484 @jihangz \n30. 0.7479 @whurobin \n31. 0.7464 myself\n32. 0.7451 @ufownl \n33. 0.7450 @gaolijing \n34. 0.7438 @truonghoang \n35. 0.7434 @zhaonat \n36. 0.7428 @alanchen1115 and team\n37. 0.7398 @cswwp347724 \n38. 0.7384 @manojprabhaakr \n39. 0.7382 @thomasdic2000 \n40. 0.7379 @sistmrl \n\nThis is obviously an incomplete list and I have undoubtedly missed many of you. If you'd like to share your highest compliant LB score, just comment below and I will try to keep the list up to date!",
      "votes": 84
    },
    {
      "id": 936089,
      "postDate": "2020-07-20T01:00:53.117Z",
      "content": "<p>0.7690 EfficientDet</p>",
      "rawMarkdown": "0.7690 EfficientDet",
      "votes": 11,
      "replies": [
        {
          "id": 939955,
          "postDate": "2020-07-22T15:33:10.300Z",
          "content": "<p>Did you get this with Pseudolabeling or without ?</p>",
          "rawMarkdown": "Did you get this with Pseudolabeling or without ?"
        },
        {
          "id": 939961,
          "postDate": "2020-07-22T15:41:18.340Z",
          "content": "<p>with pseudolabeling</p>",
          "rawMarkdown": "with pseudolabeling"
        },
        {
          "id": 943125,
          "postDate": "2020-07-24T06:48:27.613Z",
          "content": "<p><a href=\"/nguyenbadung\">@nguyenbadung</a>  Could u tell me the score without pseudo labeling?😄 Thanks</p>",
          "rawMarkdown": "@nguyenbadung  Could u tell me the score without pseudo labeling?😄 Thanks"
        },
        {
          "id": 943153,
          "postDate": "2020-07-24T07:04:04.757Z",
          "rawMarkdown": "",
          "isDeleted": true
        },
        {
          "id": 943162,
          "postDate": "2020-07-24T07:12:49.880Z",
          "rawMarkdown": "",
          "isDeleted": true
        },
        {
          "id": 947042,
          "postDate": "2020-07-27T03:28:23.483Z",
          "content": "<p>My EfficientDet based solution is at 0.765 without pseudo-labeling.</p>",
          "rawMarkdown": "My EfficientDet based solution is at 0.765 without pseudo-labeling.",
          "votes": 8
        },
        {
          "id": 951739,
          "postDate": "2020-07-30T11:03:42.573Z",
          "content": "<p>Dapeng Lee, 0.7629 efficientdet without pseudolabeling</p>",
          "rawMarkdown": "Dapeng Lee, 0.7629 efficientdet without pseudolabeling",
          "votes": 1
        },
        {
          "id": 951853,
          "postDate": "2020-07-30T13:01:35.187Z",
          "content": "<p><a href=\"/nguyenbadung\">@nguyenbadung</a> , <a href=\"/alexanderliao\">@alexanderliao</a>  Thanks!😄 our team's efficientdet have no improve with pseudolabeling ,maybe you guys find some tricks? haha</p>",
          "rawMarkdown": "@nguyenbadung , @alexanderliao  Thanks!😄 our team's efficientdet have no improve with pseudolabeling ,maybe you guys find some tricks? haha"
        }
      ]
    },
    {
      "id": 958123,
      "postDate": "2020-08-04T19:39:04.073Z",
      "content": "<p>Alright everyone, I've made my final update to the compliant leaderboard (hopefully there's no errors)! It's certainly grown a lot bigger than I initially thought it would be so I appreciate everyone who contributed their score. This competition has been an interesting  learning experience for many of us. Thank you to everyone who made this competition fun and enjoyable despite the setbacks and issues with licensing. And best of luck to all on the final leaderboard standings!</p>",
      "rawMarkdown": "Alright everyone, I've made my final update to the compliant leaderboard (hopefully there's no errors)! It's certainly grown a lot bigger than I initially thought it would be so I appreciate everyone who contributed their score. This competition has been an interesting  learning experience for many of us. Thank you to everyone who made this competition fun and enjoyable despite the setbacks and issues with licensing. And best of luck to all on the final leaderboard standings!",
      "votes": 10,
      "replies": [
        {
          "id": 958146,
          "postDate": "2020-08-04T19:54:01.403Z",
          "content": "<p>Thank you so much! It's been quite the journey :) This private LB should be interesting. </p>",
          "rawMarkdown": "Thank you so much! It's been quite the journey :) This private LB should be interesting. ",
          "votes": 1
        },
        {
          "id": 958157,
          "postDate": "2020-08-04T19:59:58.590Z",
          "content": "<p>Thanks for this. I've been following along even though I haven't submitted much. Nice to know where I stand amongst everyone listed here</p>",
          "rawMarkdown": "Thanks for this. I've been following along even though I haven't submitted much. Nice to know where I stand amongst everyone listed here",
          "votes": 2
        },
        {
          "id": 958175,
          "postDate": "2020-08-04T20:16:10.300Z",
          "content": "<p>Thanks Matthew for the original list and for all the painstaking updates!</p>",
          "rawMarkdown": "Thanks Matthew for the original list and for all the painstaking updates!",
          "votes": 3
        },
        {
          "id": 958185,
          "postDate": "2020-08-04T20:29:49.833Z",
          "content": "<p>Thanks Matthew for putting in a lot of effort. </p>",
          "rawMarkdown": "Thanks Matthew for putting in a lot of effort. ",
          "votes": 3
        },
        {
          "id": 958263,
          "postDate": "2020-08-04T22:06:45.987Z",
          "content": "<p>👍 </p>",
          "rawMarkdown": "👍 "
        }
      ]
    },
    {
      "id": 957287,
      "postDate": "2020-08-04T07:45:40.457Z",
      "content": "<p>Possibly my last submission: 0.7733 with EfficientDet-based model.</p>\n\n<p>Honestly even without all the YoloV5 subs there's still gonna be a huge shakeup. The exact settings will give out a score that varies within the 0.003 range..... Which translates to currently a gold medal and something ranking 60+.</p>",
      "rawMarkdown": "Possibly my last submission: 0.7733 with EfficientDet-based model.\n\nHonestly even without all the YoloV5 subs there's still gonna be a huge shakeup. The exact settings will give out a score that varies within the 0.003 range..... Which translates to currently a gold medal and something ranking 60+.",
      "votes": 7,
      "replies": [
        {
          "id": 957309,
          "postDate": "2020-08-04T08:01:57.707Z",
          "content": "<p>Is this using a single model?</p>",
          "rawMarkdown": "Is this using a single model?"
        }
      ]
    },
    {
      "id": 957959,
      "postDate": "2020-08-04T16:47:39.813Z",
      "content": "<p>Our mmdet-based score: 0.7736</p>",
      "rawMarkdown": "Our mmdet-based score: 0.7736",
      "votes": 8
    },
    {
      "id": 953259,
      "postDate": "2020-07-31T16:25:47.963Z",
      "content": "<p>Before deadline : 7638 [latest updated 7662] w/ FRCNN ResNest :) \n<strong>Edit</strong> answer <a href=\"/nvnnghia\">@nvnnghia</a> below, it's with pseudo labeling (thanks to your contributions). <a href=\"/gengzl\">@gengzl</a> As deadline approaces, we cannot share details, just the main idea that we use the best public ResNest, followed by <a href=\"/nvnnghia\">@nvnnghia</a> pseudo labeling framework with some adaptations to the dataset </p>",
      "rawMarkdown": "Before deadline : 7638 [latest updated 7662] w/ FRCNN ResNest :) \n**Edit** answer @nvnnghia below, it's with pseudo labeling (thanks to your contributions). @gengzl As deadline approaces, we cannot share details, just the main idea that we use the best public ResNest, followed by @nvnnghia pseudo labeling framework with some adaptations to the dataset ",
      "votes": 5,
      "replies": [
        {
          "id": 953315,
          "postDate": "2020-07-31T17:22:41.327Z",
          "content": "<p>In addition to this I think <a href=\"/kyoshioka47\">@kyoshioka47</a>  score can be removed....\nWe merged..\nNot to cause any confusion 😅 </p>",
          "rawMarkdown": "In addition to this I think @kyoshioka47  score can be removed....\nWe merged..\nNot to cause any confusion 😅 ",
          "votes": 2
        },
        {
          "id": 953596,
          "postDate": "2020-07-31T22:49:09.283Z",
          "content": "<p>very impressive!! Is that with or without pseudo labeling</p>",
          "rawMarkdown": "very impressive!! Is that with or without pseudo labeling",
          "votes": 2
        },
        {
          "id": 953672,
          "postDate": "2020-08-01T01:37:24.983Z",
          "rawMarkdown": "",
          "isDeleted": true
        }
      ]
    },
    {
      "id": 941469,
      "postDate": "2020-07-23T08:27:39.053Z",
      "content": "<p>0.7700.(Effdet)</p>",
      "rawMarkdown": "0.7700.(Effdet)",
      "votes": 5,
      "replies": [
        {
          "id": 944903,
          "postDate": "2020-07-25T12:45:35.007Z",
          "content": "<p>Update: 0.7730(Effdet)</p>",
          "rawMarkdown": "Update: 0.7730(Effdet)",
          "votes": 5
        },
        {
          "id": 947935,
          "postDate": "2020-07-27T15:03:29.447Z",
          "content": "<p>which version effdet you used?</p>",
          "rawMarkdown": "which version effdet you used?\n"
        },
        {
          "id": 953588,
          "postDate": "2020-07-31T22:36:11.097Z",
          "content": "<p>Update: 0.7753(Effdet)</p>\n\n<p>&gt; which version effdet you used?  </p>\n\n<p>It's secret until over deadline.  </p>",
          "rawMarkdown": "Update: 0.7753(Effdet)\n\n&gt; which version effdet you used?  \n  \nIt's secret until over deadline.  ",
          "votes": 3
        }
      ]
    },
    {
      "id": 953614,
      "postDate": "2020-07-31T23:16:22.477Z",
      "content": "<p>0.7514 - Retinanet - With EfficientNet_backbone :)</p>",
      "rawMarkdown": "0.7514 - Retinanet - With EfficientNet_backbone :)",
      "votes": 3
    },
    {
      "id": 949647,
      "postDate": "2020-07-28T19:09:45.867Z",
      "content": "<p>0.7484 CenterNet, 8:2 single split, TTAx3 (fliplr &amp; flipud), no pseudo label\n[Update] 0.7520 with TTAx9\n[Update 2] 0.7588 with 3 folds &amp;, TTAx6\n[Final Update] 0.7590 with 3 folds, TTAx6; no luck with pseudo label ☹️ </p>",
      "rawMarkdown": "0.7484 CenterNet, 8:2 single split, TTAx3 (fliplr &amp; flipud), no pseudo label\n[Update] 0.7520 with TTAx9\n[Update 2] 0.7588 with 3 folds &amp;, TTAx6\n[Final Update] 0.7590 with 3 folds, TTAx6; no luck with pseudo label ☹️ ",
      "votes": 3,
      "replies": [
        {
          "id": 951624,
          "postDate": "2020-07-30T09:15:06.243Z",
          "content": "<p>what's the meaning of TTAx9?</p>",
          "rawMarkdown": "what's the meaning of TTAx9?"
        },
        {
          "id": 952026,
          "postDate": "2020-07-30T15:08:53.103Z",
          "content": "<p>TTA is test time augmentation, I think 9 is 9 augmentations per photo (but maybe not)</p>",
          "rawMarkdown": "TTA is test time augmentation, I think 9 is 9 augmentations per photo (but maybe not)"
        },
        {
          "id": 953503,
          "postDate": "2020-07-31T20:54:55.897Z",
          "content": "<p>Original, flip, multi-scale. 8 augmentations per image + original</p>",
          "rawMarkdown": "Original, flip, multi-scale. 8 augmentations per image + original"
        },
        {
          "id": 953627,
          "postDate": "2020-07-31T23:38:52.290Z",
          "content": "<p>Thank you for your sharing👍 </p>",
          "rawMarkdown": "Thank you for your sharing👍 "
        }
      ]
    },
    {
      "id": 947926,
      "postDate": "2020-07-27T15:00:20.507Z",
      "content": "<p>0.7573 mmdetection models \n[Update] 0.7601</p>",
      "rawMarkdown": "0.7573 mmdetection models \n[Update] 0.7601",
      "votes": 3
    },
    {
      "id": 930433,
      "postDate": "2020-07-15T13:06:40.203Z",
      "content": "<p>I have fastercnn with 0.7491</p>",
      "rawMarkdown": "I have fastercnn with 0.7491",
      "votes": 3,
      "replies": [
        {
          "id": 930859,
          "postDate": "2020-07-15T18:52:36.067Z",
          "content": "<p>Added!</p>",
          "rawMarkdown": "Added!"
        }
      ]
    },
    {
      "id": 929903,
      "postDate": "2020-07-15T04:11:13.183Z",
      "content": "<p>0.7601 EfficientDet</p>",
      "rawMarkdown": "0.7601 EfficientDet",
      "votes": 3,
      "replies": [
        {
          "id": 929923,
          "postDate": "2020-07-15T04:46:15.260Z",
          "content": "<p>Thanks, added to the list!</p>",
          "rawMarkdown": "Thanks, added to the list!"
        }
      ]
    },
    {
      "id": 957130,
      "postDate": "2020-08-04T04:48:59.937Z",
      "content": "<p>0.7560 EffdetD5 </p>",
      "rawMarkdown": "0.7560 EffdetD5 ",
      "votes": 1
    },
    {
      "id": 957091,
      "postDate": "2020-08-04T03:39:08.547Z",
      "content": "<p>0.7597 with effdet d5</p>",
      "rawMarkdown": "0.7597 with effdet d5",
      "votes": 1
    },
    {
      "id": 956477,
      "postDate": "2020-08-03T14:32:32.167Z",
      "content": "<p>I'm compliant at 0.7438</p>",
      "rawMarkdown": "I'm compliant at 0.7438",
      "votes": 1
    },
    {
      "id": 954849,
      "postDate": "2020-08-02T06:21:45.473Z",
      "content": "<p>[UPDATE] 0.7602 using EfficientDet5 backbone.</p>",
      "rawMarkdown": "[UPDATE] 0.7602 using EfficientDet5 backbone.",
      "votes": 1,
      "replies": [
        {
          "id": 954853,
          "postDate": "2020-08-02T06:28:23.037Z",
          "content": "<p>Is this with pseudo-labeling??</p>",
          "rawMarkdown": "Is this with pseudo-labeling??"
        },
        {
          "id": 954856,
          "postDate": "2020-08-02T06:32:50.440Z",
          "content": "<p>Yes, with Pseduo-Labelling.</p>",
          "rawMarkdown": "Yes, with Pseduo-Labelling.",
          "votes": 1
        },
        {
          "id": 954870,
          "postDate": "2020-08-02T06:48:40Z",
          "content": "<p><a href=\"/watzisname\">@watzisname</a> oh brother, so this means your current best score is yolov5 but managed to get 0.7575 on effdet?</p>",
          "rawMarkdown": "@watzisname oh brother, so this means your current best score is yolov5 but managed to get 0.7575 on effdet?"
        },
        {
          "id": 955128,
          "postDate": "2020-08-02T11:22:32.547Z",
          "content": "<p><a href=\"/reighns\">@reighns</a>\nYes that is correct.</p>",
          "rawMarkdown": "@reighns\nYes that is correct."
        }
      ]
    },
    {
      "id": 951192,
      "postDate": "2020-07-30T01:24:04.260Z",
      "content": "<p>0.7496 effdet, 0.7115 yolov4 without PL or OOF. Thanks a ton Matthew!</p>",
      "rawMarkdown": "0.7496 effdet, 0.7115 yolov4 without PL or OOF. Thanks a ton Matthew!",
      "votes": 1
    },
    {
      "id": 950889,
      "postDate": "2020-07-29T17:34:13.497Z",
      "content": "<p><a href=\"/matthewmasters\">@matthewmasters</a> - I have been following your posts in this competition. You have been doing an amazing work. 0.7378 with EfficientDet\nUpdate: 0.7384 EfficientDet</p>",
      "rawMarkdown": "@matthewmasters - I have been following your posts in this competition. You have been doing an amazing work. 0.7378 with EfficientDet\nUpdate: 0.7384 EfficientDet",
      "votes": 1,
      "replies": [
        {
          "id": 951049,
          "postDate": "2020-07-29T20:48:30.450Z",
          "content": "<p>Thanks for the kind words. I added your score to the LB</p>",
          "rawMarkdown": "Thanks for the kind words. I added your score to the LB",
          "votes": 1
        }
      ]
    },
    {
      "id": 950264,
      "postDate": "2020-07-29T09:57:49.770Z",
      "content": "<p>0.7450 with EfficientDet</p>",
      "rawMarkdown": "0.7450 with EfficientDet",
      "votes": 1
    },
    {
      "id": 945222,
      "postDate": "2020-07-25T16:57:44.033Z",
      "content": "<p>Sorry for the delay in updates. I have just updated the list again</p>",
      "rawMarkdown": "Sorry for the delay in updates. I have just updated the list again",
      "votes": 1
    },
    {
      "id": 935574,
      "postDate": "2020-07-19T13:55:50.440Z",
      "content": "<p>YoloV3 0.7509, my compliant entry.</p>",
      "rawMarkdown": "YoloV3 0.7509, my compliant entry.",
      "votes": 1,
      "replies": [
        {
          "id": 951183,
          "postDate": "2020-07-30T01:15:36.083Z",
          "rawMarkdown": "",
          "isDeleted": true
        }
      ]
    },
    {
      "id": 934969,
      "postDate": "2020-07-19T01:01:50.260Z",
      "content": "<p><a href=\"/nvnnghia\">@nvnnghia</a> Very nice work! I updated the listing</p>\n\n<p>(not sure why but it wouldn't let me reply directly to your comment)</p>",
      "rawMarkdown": "@nvnnghia Very nice work! I updated the listing\n\n(not sure why but it wouldn't let me reply directly to your comment)",
      "votes": 1
    },
    {
      "id": 934958,
      "postDate": "2020-07-19T00:34:26.610Z",
      "content": "<p>[UPDATED] 0.7683</p>",
      "rawMarkdown": "[UPDATED] 0.7683\n",
      "votes": 1
    },
    {
      "id": 932400,
      "postDate": "2020-07-17T03:29:22.220Z",
      "content": "<p>EfficientDet (no psuedo) 0.7428, alan mise team's compliant entry.</p>",
      "rawMarkdown": "EfficientDet (no psuedo) 0.7428, alan mise team's compliant entry.",
      "votes": 1,
      "replies": [
        {
          "id": 932679,
          "postDate": "2020-07-17T07:46:50.560Z",
          "content": "<p>Forgot to respond but I added this earlier! :)</p>",
          "rawMarkdown": "Forgot to respond but I added this earlier! :)",
          "votes": 2
        }
      ]
    },
    {
      "id": 932362,
      "postDate": "2020-07-17T02:31:31.220Z",
      "content": "<p>EfficientDet: 0.7365</p>",
      "rawMarkdown": "EfficientDet: 0.7365",
      "votes": 1,
      "replies": [
        {
          "id": 932364,
          "postDate": "2020-07-17T02:34:54.987Z",
          "content": "<p>Added</p>",
          "rawMarkdown": "Added",
          "votes": 1
        }
      ]
    },
    {
      "id": 931935,
      "postDate": "2020-07-16T15:06:37.520Z",
      "content": "<p>EfficientDet: 0.7382; My compliant entry.</p>",
      "rawMarkdown": "EfficientDet: 0.7382; My compliant entry.",
      "votes": 1,
      "replies": [
        {
          "id": 932099,
          "postDate": "2020-07-16T17:37:54.747Z",
          "content": "<p>Added!</p>",
          "rawMarkdown": "Added!"
        }
      ]
    },
    {
      "id": 931759,
      "postDate": "2020-07-16T12:30:52.310Z",
      "content": "<p>0.7510 EfficientDet\n[update]0.7582</p>",
      "rawMarkdown": "0.7510 EfficientDet\n[update]0.7582",
      "votes": 1,
      "replies": [
        {
          "id": 932100,
          "postDate": "2020-07-16T17:38:02.170Z",
          "content": "<p>Added!</p>",
          "rawMarkdown": "Added!"
        },
        {
          "id": 932674,
          "postDate": "2020-07-17T07:40:19.547Z",
          "content": "<p>Thanks <a href=\"/matthewmasters\">@matthewmasters</a> . Maybe there should be a positive contribution award for people like you who contribute to the community</p>",
          "rawMarkdown": "Thanks @matthewmasters . Maybe there should be a positive contribution award for people like you who contribute to the community",
          "votes": 3
        },
        {
          "id": 932678,
          "postDate": "2020-07-17T07:46:26.980Z",
          "content": "<p>Well I'm glad you appreciate my contribution, that means a lot! Actually this morning I got an email from a member of the Kaggle team to send me swag for my active participation, so that's was certainly nice to receive! 😀 </p>",
          "rawMarkdown": "Well I'm glad you appreciate my contribution, that means a lot! Actually this morning I got an email from a member of the Kaggle team to send me swag for my active participation, so that's was certainly nice to receive! 😀 ",
          "votes": 2
        },
        {
          "id": 932685,
          "postDate": "2020-07-17T07:49:08.863Z",
          "content": "<p><a href=\"/guizengyou\">@guizengyou</a> Totally agreed \n<a href=\"/matthewmasters\">@matthewmasters</a>  I am really glad to hear that , you deserve it bro</p>",
          "rawMarkdown": "@guizengyou Totally agreed \n@matthewmasters  I am really glad to hear that , you deserve it bro",
          "votes": 1
        },
        {
          "id": 947893,
          "postDate": "2020-07-27T14:41:16.877Z",
          "content": "<p>Did you use D5 or D7</p>",
          "rawMarkdown": "Did you use D5 or D7"
        }
      ]
    },
    {
      "id": 928816,
      "postDate": "2020-07-14T08:35:10.700Z",
      "content": "<p>EfficientDet ( I think it`s ok ) 0.7548</p>\n\n<p>Update: 0.7614 without PseudoLabeling</p>",
      "rawMarkdown": "EfficientDet ( I think it`s ok ) 0.7548\n\nUpdate: 0.7614 without PseudoLabeling",
      "votes": 1,
      "replies": [
        {
          "id": 928820,
          "postDate": "2020-07-14T08:37:42.520Z",
          "content": "<p>Thanks, added! PS. EfficientDet is okay but has some specific requirements. The creator of the pytorch EfficientDet implementation can be found <a href=\"https://www.kaggle.com/c/global-wheat-detection/discussion/164845#923524\">here</a> giving some advice to this end.</p>",
          "rawMarkdown": "Thanks, added! PS. EfficientDet is okay but has some specific requirements. The creator of the pytorch EfficientDet implementation can be found [here](https://www.kaggle.com/c/global-wheat-detection/discussion/164845#923524) giving some advice to this end.",
          "votes": 1
        },
        {
          "id": 928823,
          "postDate": "2020-07-14T08:41:00.390Z",
          "content": "<p>Thanks :) Good to know I`ll add this.</p>\n\n<p>BTW, can someone please write what models are ok ? Because right now it`s hard to catch up with all the informations</p>",
          "rawMarkdown": "Thanks :) Good to know I`ll add this.\n\nBTW, can someone please write what models are ok ? Because right now it`s hard to catch up with all the informations",
          "votes": 1
        },
        {
          "id": 928824,
          "postDate": "2020-07-14T08:44:38.463Z",
          "content": "<p>Yes actually! <a href=\"/nakajima\">@nakajima</a> put together <a href=\"https://www.kaggle.com/c/global-wheat-detection/discussion/165735#924609\">this list</a> which should help you with that</p>",
          "rawMarkdown": "Yes actually! @nakajima put together [this list](https://www.kaggle.com/c/global-wheat-detection/discussion/165735#924609) which should help you with that"
        }
      ]
    },
    {
      "id": 928804,
      "postDate": "2020-07-14T08:20:42.450Z",
      "content": "<p>I'm compliant at 0.7669</p>",
      "rawMarkdown": "I'm compliant at 0.7669",
      "votes": 1,
      "replies": [
        {
          "id": 928805,
          "postDate": "2020-07-14T08:22:44.030Z",
          "content": "<p>Well done! I added it to the list</p>",
          "rawMarkdown": "Well done! I added it to the list"
        }
      ]
    },
    {
      "id": 928733,
      "postDate": "2020-07-14T07:18:07.287Z",
      "content": "<p>I'm compliant at 0.7529</p>",
      "rawMarkdown": "I'm compliant at 0.7529",
      "votes": 1,
      "replies": [
        {
          "id": 928739,
          "postDate": "2020-07-14T07:20:46.517Z",
          "content": "<p>Good work! I just added you to the list :)</p>",
          "rawMarkdown": "Good work! I just added you to the list :)",
          "votes": 1
        }
      ]
    },
    {
      "id": 956777,
      "postDate": "2020-08-03T19:10:07.823Z",
      "content": "<p>[UPDATE]- 0.7695 FRCNN\nfrom <a href=\"/datahobbit\">@datahobbit</a> and team...\nAlso <a href=\"/kyoshioka47\">@kyoshioka47</a>'s score can be removed as we are in same team...\nThank you</p>",
      "rawMarkdown": "[UPDATE]- 0.7695 FRCNN\nfrom @datahobbit and team...\nAlso @kyoshioka47's score can be removed as we are in same team...\nThank you",
      "votes": 2
    },
    {
      "id": 954039,
      "postDate": "2020-08-01T10:30:02.237Z",
      "content": "<p><a href=\"/matthewmasters\">@matthewmasters</a> ,Thanks ,matthew! ours compliant entry: 0.7625.</p>",
      "rawMarkdown": "@matthewmasters ,Thanks ,matthew! ours compliant entry: 0.7625.",
      "votes": 2
    },
    {
      "id": 951048,
      "postDate": "2020-07-29T20:48:09.847Z",
      "content": "<p>Updated again. Seeing as its the last week, I will likely only make one more update before the competition end</p>",
      "rawMarkdown": "Updated again. Seeing as its the last week, I will likely only make one more update before the competition end",
      "votes": 2,
      "replies": [
        {
          "id": 953429,
          "postDate": "2020-07-31T19:32:35.747Z",
          "content": "<p>Thanks Matthew. Your efforts on this have been appreciated. In a way I like the fact it isn't updated instantly like the usual leaderboard because there is less temptation to focus too heavily on leaderboard scores. At the same time, it is useful to have some benchmark to see how we are doing against non-Yolov5 scores of other teams.</p>",
          "rawMarkdown": "Thanks Matthew. Your efforts on this have been appreciated. In a way I like the fact it isn't updated instantly like the usual leaderboard because there is less temptation to focus too heavily on leaderboard scores. At the same time, it is useful to have some benchmark to see how we are doing against non-Yolov5 scores of other teams.",
          "votes": 4
        }
      ]
    },
    {
      "id": 946142,
      "postDate": "2020-07-26T11:38:56.287Z",
      "content": "<p>0.7550 Effdet + TTAx16 + pseudo labelling</p>",
      "rawMarkdown": "0.7550 Effdet + TTAx16 + pseudo labelling",
      "votes": 2
    },
    {
      "id": 946031,
      "postDate": "2020-07-26T09:51:33.177Z",
      "content": "<p>0.7541 EffDet d5 + d7x + FRCNN\n[Update] 0.7591\n[Update] 0.7613</p>",
      "rawMarkdown": "0.7541 EffDet d5 + d7x + FRCNN\n[Update] 0.7591\n[Update] 0.7613",
      "votes": 2
    },
    {
      "id": 940717,
      "postDate": "2020-07-23T05:24:41.560Z",
      "content": "<p>effdet 0.7680....</p>",
      "rawMarkdown": "effdet 0.7680....",
      "votes": 2
    },
    {
      "id": 936377,
      "postDate": "2020-07-20T07:12:20.190Z",
      "content": "<p>Updated with scores of <a href=\"/nguyenbadung\">@nguyenbadung</a>, <a href=\"/whurobin\">@whurobin</a>, and <a href=\"/tiagotml\">@tiagotml</a> </p>",
      "rawMarkdown": "Updated with scores of @nguyenbadung, @whurobin, and @tiagotml ",
      "votes": 2
    },
    {
      "id": 936369,
      "postDate": "2020-07-20T07:06:40.510Z",
      "content": "<p>0.7466\n[UPDATED] 0.7479\n[UPDATED] 0.7551</p>",
      "rawMarkdown": "0.7466\n[UPDATED] 0.7479\n[UPDATED] 0.7551",
      "votes": 2
    },
    {
      "id": 929419,
      "postDate": "2020-07-14T16:52:29.737Z",
      "content": "<p><a href=\"/matthewmasters\">@matthewmasters</a> you have been really active in this competition , I really thank you for all your discussion threads and work\nI would be glad if you clear my following doubts:-\n* If a team uses yolo will it be removed from the lb? or only the winning solutions should not use , but if you geta gold using yolo its ohk?  There is a lot of confusion around it\n* The images which the correction is done are present only in 38 percent hidden test set ? Also is there a dataset compiled by someone which contains all the images which have noisy bboxes?</p>\n\n<p>Thanks in advance</p>",
      "rawMarkdown": "@matthewmasters you have been really active in this competition , I really thank you for all your discussion threads and work\nI would be glad if you clear my following doubts:-\n* If a team uses yolo will it be removed from the lb? or only the winning solutions should not use , but if you geta gold using yolo its ohk?  There is a lot of confusion around it\n* The images which the correction is done are present only in 38 percent hidden test set ? Also is there a dataset compiled by someone which contains all the images which have noisy bboxes?\n\nThanks in advance",
      "votes": 2,
      "replies": [
        {
          "id": 929621,
          "postDate": "2020-07-14T19:52:12.893Z",
          "content": "<p>Hi there, thanks for the kind words. To answer your questions:\n1. From what the competition organizers have said (<a href=\"https://www.kaggle.com/c/global-wheat-detection/discussion/165735\">here</a> and <a href=\"https://www.kaggle.com/c/global-wheat-detection/discussion/165735#928432\">here</a>), the two final submissions you select should fit the licensing term (i.e. no YoloV5). So yes, even if you are just going for medals and not the cash prize, you must abide by the rules. If they find that you did not follow the rules, they reserve the right to disqualify you. Though they haven't been very clear on how exactly this will be enforced.\n2. Yes, the issue with large bounding boxes will only be fixed for the 38% of private test data. You can read the full response from the organizers <a href=\"https://www.kaggle.com/c/global-wheat-detection/discussion/163291\">here</a>. Some users have kindly shared some of these boxes they have found in the training set <a href=\"https://www.kaggle.com/c/global-wheat-detection/discussion/159578\">here</a>. There was also some discussion about tiny bounding boxes <a href=\"https://www.kaggle.com/c/global-wheat-detection/discussion/154774\">here</a>. The easiest way to fix these is just to filter based on box length or area. I haven't looked at them but it looks like there are <a href=\"https://www.kaggle.com/raininbox/check-clean-big-small-bboxes\">some helpful notebooks</a> where people have already done this.</p>",
          "rawMarkdown": "Hi there, thanks for the kind words. To answer your questions:\n1. From what the competition organizers have said ([here](https://www.kaggle.com/c/global-wheat-detection/discussion/165735) and [here](https://www.kaggle.com/c/global-wheat-detection/discussion/165735#928432)), the two final submissions you select should fit the licensing term (i.e. no YoloV5). So yes, even if you are just going for medals and not the cash prize, you must abide by the rules. If they find that you did not follow the rules, they reserve the right to disqualify you. Though they haven't been very clear on how exactly this will be enforced.\n2. Yes, the issue with large bounding boxes will only be fixed for the 38% of private test data. You can read the full response from the organizers [here](https://www.kaggle.com/c/global-wheat-detection/discussion/163291). Some users have kindly shared some of these boxes they have found in the training set [here](https://www.kaggle.com/c/global-wheat-detection/discussion/159578). There was also some discussion about tiny bounding boxes [here](https://www.kaggle.com/c/global-wheat-detection/discussion/154774). The easiest way to fix these is just to filter based on box length or area. I haven't looked at them but it looks like there are [some helpful notebooks](https://www.kaggle.com/raininbox/check-clean-big-small-bboxes) where people have already done this.",
          "votes": 2
        },
        {
          "id": 929657,
          "postDate": "2020-07-14T20:27:45.413Z",
          "content": "<p>Thanks a lot for the detailed response</p>",
          "rawMarkdown": "Thanks a lot for the detailed response",
          "votes": 1
        },
        {
          "id": 929679,
          "postDate": "2020-07-14T20:58:45.950Z",
          "content": "<p>Dear <a href=\"/matthewmasters\">@matthewmasters</a>, even if I value your word highly, I am not sure if you are correct.</p>\n\n<p>There are only 3 prices as you can see here:\n<a href=\"https://www.kaggle.com/c/global-wheat-detection/overview/prizes\">https://www.kaggle.com/c/global-wheat-detection/overview/prizes</a></p>\n\n<p>And Kaggle team stated [here ] that it's not possible to win PRIZES using yolov5, they didn't mention Medals.</p>\n\n<p>(<a href=\"https://www.kaggle.com/c/global-wheat-detection/discussion/163433)on\">https://www.kaggle.com/c/global-wheat-detection/discussion/163433)on</a> 14.07.2020 (just an hour ago)\n<img src=\"https://www.googleapis.com/download/storage/v1/b/kaggle-user-content/o/inbox%2F5006160%2Fe88f582667224fbd39ddf44e5f45ae78%2FUnbenannt.PNG?generation=1594760110608297&amp;alt=media\" alt=\"\"></p>\n\n<p>Please don't get me wrong, it's not about fairness - I think we all should be fair, of course.\nIt's about clarity and it still feels like we still don't have that. It feels like we are trying to get part-time-lawyers interpreting the few words said by the Host and the Kaggle team.\nSo, what's coming? A huge shake-up, as 500+ teams/solutions using yolov5 will be deleted? Or just manual checking of the first 3 solutions for the prices? I only have my guess.</p>\n\n<p>I don't want to complain (sorry if I did), but i want strongly suggest more clarity to get the focus back from <em>\"what if I use model/license xzy...\"</em> back to <em>\"what if I modify the wbf ...\"</em>\n<a href=\"/tanulsingh077\">@tanulsingh077</a>, sorry for not solving this ;)</p>",
          "rawMarkdown": "Dear @matthewmasters, even if I value your word highly, I am not sure if you are correct.\n\nThere are only 3 prices as you can see here:\n[https://www.kaggle.com/c/global-wheat-detection/overview/prizes](https://www.kaggle.com/c/global-wheat-detection/overview/prizes)\n\nAnd Kaggle team stated [here ] that it's not possible to win PRIZES using yolov5, they didn't mention Medals.\n\n(https://www.kaggle.com/c/global-wheat-detection/discussion/163433)on 14.07.2020 (just an hour ago)\n![](https://www.googleapis.com/download/storage/v1/b/kaggle-user-content/o/inbox%2F5006160%2Fe88f582667224fbd39ddf44e5f45ae78%2FUnbenannt.PNG?generation=1594760110608297&amp;alt=media)\n\nPlease don't get me wrong, it's not about fairness - I think we all should be fair, of course.\nIt's about clarity and it still feels like we still don't have that. It feels like we are trying to get part-time-lawyers interpreting the few words said by the Host and the Kaggle team.\nSo, what's coming? A huge shake-up, as 500+ teams/solutions using yolov5 will be deleted? Or just manual checking of the first 3 solutions for the prices? I only have my guess.\n\nI don't want to complain (sorry if I did), but i want strongly suggest more clarity to get the focus back from *\"what if I use model/license xzy...\"* back to *\"what if I modify the wbf ...\"*\n@tanulsingh077, sorry for not solving this ;)\n\n\n",
          "votes": 8
        },
        {
          "id": 929696,
          "postDate": "2020-07-14T21:25:59.300Z",
          "content": "<p>I agree that the Kaggle team and competition organizers have been very vague and unclear in their responses. Like in the above screenshot, what does \"prizes\" mean exactly? Just the cash prizes or are medals included?</p>\n\n<p>I made my response above based on several of the posts made by <a href=\"/maggiemd\">@maggiemd</a> and <a href=\"/stavness\">@stavness</a> which I will outline below (I apologize for the long post, I'm just trying to include all the relevant responses).</p>\n\n<ul>\n<li>In the <a href=\"https://www.kaggle.com/c/global-wheat-detection/discussion/165735\">\"Competition Rules\" thread</a>, Ian said:\n&gt; You are free to try out any technique you like while crafting your submissions; however the two submissions that you select as your final submissions must be compliant with the competition rules.</li>\n</ul>\n\n<p>This indicates to me that even if you're going for medals, you must abide by the rules. (As medals are awarded based on the two submissions you select for final scoring).</p>\n\n<ul>\n<li><p>In response to that post, <a href=\"https://www.kaggle.com/c/global-wheat-detection/discussion/165735#923546\">I tried to clear up how exactly the rules would be enforced</a> but the response I got from Maggie simply reiterated the previous point:\n&gt; The requirement is that the two final submissions must comply with the rules. Individual submissions can be submitted throughout but the two selected must conform.</p></li>\n<li><p>In the same thread, <a href=\"https://www.kaggle.com/c/global-wheat-detection/discussion/165735#923577\">another user tried to clarify the same point</a>. Maggie responded with:\n&gt; As with any competition, if we become aware that someone had not conformed with the rules then the host has the discretion to disqualify.</p></li>\n<li><p>Finally, I consulted the rules for any mention of \"medals\" and found the following:\n&gt; A disqualified participant may be removed from the Competition leaderboard, at Kaggle's sole discretion. If a Participant is removed from the Competition Leaderboard, additional winning features associated with the Kaggle competition platform, for example Kaggle points or medals, may also not be awarded.</p></li>\n</ul>\n\n<p>Therefor, I came to the conclusion I stated above, that even solutions going for medals must abide by the licensing rule. Of course I am not 100% certain of this since the posts from Kaggle themselves is unclear and this is just my interpretation of what they have said. In order to be certain, it would be nice if they could actually answer the questions being asked instead of reiterating the same point.</p>",
          "rawMarkdown": "I agree that the Kaggle team and competition organizers have been very vague and unclear in their responses. Like in the above screenshot, what does \"prizes\" mean exactly? Just the cash prizes or are medals included?\n\nI made my response above based on several of the posts made by @maggiemd and @stavness which I will outline below (I apologize for the long post, I'm just trying to include all the relevant responses).\n\n- In the [\"Competition Rules\" thread](https://www.kaggle.com/c/global-wheat-detection/discussion/165735), Ian said:\n&gt; You are free to try out any technique you like while crafting your submissions; however the two submissions that you select as your final submissions must be compliant with the competition rules.\n\nThis indicates to me that even if you're going for medals, you must abide by the rules. (As medals are awarded based on the two submissions you select for final scoring).\n\n- In response to that post, [I tried to clear up how exactly the rules would be enforced](https://www.kaggle.com/c/global-wheat-detection/discussion/165735#923546) but the response I got from Maggie simply reiterated the previous point:\n&gt; The requirement is that the two final submissions must comply with the rules. Individual submissions can be submitted throughout but the two selected must conform.\n\n- In the same thread, [another user tried to clarify the same point](https://www.kaggle.com/c/global-wheat-detection/discussion/165735#923577). Maggie responded with:\n&gt; As with any competition, if we become aware that someone had not conformed with the rules then the host has the discretion to disqualify.\n\n- Finally, I consulted the rules for any mention of \"medals\" and found the following:\n&gt; A disqualified participant may be removed from the Competition leaderboard, at Kaggle's sole discretion. If a Participant is removed from the Competition Leaderboard, additional winning features associated with the Kaggle competition platform, for example Kaggle points or medals, may also not be awarded.\n\nTherefor, I came to the conclusion I stated above, that even solutions going for medals must abide by the licensing rule. Of course I am not 100% certain of this since the posts from Kaggle themselves is unclear and this is just my interpretation of what they have said. In order to be certain, it would be nice if they could actually answer the questions being asked instead of reiterating the same point.",
          "votes": 8
        },
        {
          "id": 929773,
          "postDate": "2020-07-14T23:48:52.147Z",
          "content": "<p>I guess Kaggle team cannot be too bold (state clearly that they will check all invalid) since they do not have enough man-power .. But they will do their best,for verifying the top rankers</p>",
          "rawMarkdown": "I guess Kaggle team cannot be too bold (state clearly that they will check all invalid) since they do not have enough man-power .. But they will do their best,for verifying the top rankers"
        },
        {
          "id": 929909,
          "postDate": "2020-07-15T04:17:44.970Z",
          "content": "<p><a href=\"/ratthachat\">@ratthachat</a> I totally agree that kaggle always verifies the top ranker. But there are two options here either use Yolo and end up in silver or bronze lb or use effdet or another model that comply with the rules and end up outside of medal zone. So in the case of non removal of Yolo using teams, the teams which tried to comply with the rules might end up with no medals and their time and effort will be wasted. Kaggle always does their best, it would be highly appreciable to get some clarity in this matter too right?\nThe question again comes down to \nWhat should we do when we are going just for the medals, should be comply pr not\nThanks</p>",
          "rawMarkdown": "@ratthachat I totally agree that kaggle always verifies the top ranker. But there are two options here either use Yolo and end up in silver or bronze lb or use effdet or another model that comply with the rules and end up outside of medal zone. So in the case of non removal of Yolo using teams, the teams which tried to comply with the rules might end up with no medals and their time and effort will be wasted. Kaggle always does their best, it would be highly appreciable to get some clarity in this matter too right?\nThe question again comes down to \nWhat should we do when we are going just for the medals, should be comply pr not\nThanks",
          "votes": 1
        },
        {
          "id": 930270,
          "postDate": "2020-07-15T10:10:28.967Z",
          "content": "<p>Hi <a href=\"/tanulsingh077\">@tanulsingh077</a> , to be honest, we all have to choose our own path and accept the risk we choose. </p>\n\n<p>To our team, we truly believe the correct way is to comply the rule by not using YoloV5.</p>",
          "rawMarkdown": "Hi @tanulsingh077 , to be honest, we all have to choose our own path and accept the risk we choose. \n\nTo our team, we truly believe the correct way is to comply the rule by not using YoloV5.",
          "votes": 4
        },
        {
          "id": 930302,
          "postDate": "2020-07-15T10:49:18.480Z",
          "content": "<p>Agree.\nI have yolo 7745 but I'm not using it  because its not ok.</p>\n\n<p>As you said , its everyone choise but most ppl will choose use yolo and get high rank , \nto get gold medal and put it into cv: kaggle gold medal top 0.00000001 world oh yeah !!</p>",
          "rawMarkdown": "Agree.\nI have yolo 7745 but I'm not using it  because its not ok.\n\nAs you said , its everyone choise but most ppl will choose use yolo and get high rank , \nto get gold medal and put it into cv: kaggle gold medal top 0.00000001 world oh yeah !!",
          "votes": 3
        },
        {
          "id": 930862,
          "postDate": "2020-07-15T18:54:49.573Z",
          "content": "<p><a href=\"/chrisden\">@chrisden</a> <a href=\"/tanulsingh077\">@tanulsingh077</a> \nWell I suppose we have <a href=\"https://www.kaggle.com/c/global-wheat-detection/discussion/167140#930712\">our response</a>. Albeit an unprofessional one.</p>",
          "rawMarkdown": "@chrisden @tanulsingh077 \nWell I suppose we have [our response](https://www.kaggle.com/c/global-wheat-detection/discussion/167140#930712). Albeit an unprofessional one.",
          "votes": 1
        },
        {
          "id": 930895,
          "postDate": "2020-07-15T19:45:52.700Z",
          "content": "<p>Thanks, appreciated, and fully agree on the level of professionality of this answer ;)</p>",
          "rawMarkdown": "Thanks, appreciated, and fully agree on the level of professionality of this answer ;)",
          "votes": 1
        },
        {
          "id": 945574,
          "postDate": "2020-07-26T01:16:27.363Z",
          "content": "<p><a href=\"/tanulsingh077\">@tanulsingh077</a>, I think the right answer is yes all teams with Yolov5 submission chosen as one of their final 2 subs will be removed from the LB. I do not understand why people are confused. It is not hard to write a simple script that goes through 4000+ submissions to check for Yolov5 models hence I believe Kaggle will do it. Heck if this were posted as Kaggle challenge, many teams would score 99.99 accuracy within the 1st week. Just my 2 cents :-)</p>",
          "rawMarkdown": "@tanulsingh077, I think the right answer is yes all teams with Yolov5 submission chosen as one of their final 2 subs will be removed from the LB. I do not understand why people are confused. It is not hard to write a simple script that goes through 4000+ submissions to check for Yolov5 models hence I believe Kaggle will do it. Heck if this were posted as Kaggle challenge, many teams would score 99.99 accuracy within the 1st week. Just my 2 cents :-)",
          "votes": 2
        }
      ]
    },
    {
      "id": 928828,
      "postDate": "2020-07-14T08:47:15.883Z",
      "content": "<p>I'm compliant at 0.7613 :)</p>",
      "rawMarkdown": "I'm compliant at 0.7613 :)",
      "votes": 2,
      "replies": [
        {
          "id": 928834,
          "postDate": "2020-07-14T08:50:42.963Z",
          "content": "<p>Thanks and well done! Added it</p>",
          "rawMarkdown": "Thanks and well done! Added it"
        }
      ]
    },
    {
      "id": 928716,
      "postDate": "2020-07-14T07:10:41.380Z",
      "content": "<p>.7536 mostly compliant apart from wbf. But I'm sure everyone has that problem and if they feel they are going to win it they'll just write their own implementation.</p>",
      "rawMarkdown": ".7536 mostly compliant apart from wbf. But I'm sure everyone has that problem and if they feel they are going to win it they'll just write their own implementation.",
      "votes": 2,
      "replies": [
        {
          "id": 928722,
          "postDate": "2020-07-14T07:13:56.860Z",
          "content": "<p>Added! Thank you for sharing :)</p>",
          "rawMarkdown": "Added! Thank you for sharing :)"
        },
        {
          "id": 929017,
          "postDate": "2020-07-14T11:59:23.270Z",
          "content": "<p><a href=\"/alexandersoare\">@alexandersoare</a> I could not find a licence on WBF, did you find it ?</p>",
          "rawMarkdown": "@alexandersoare I could not find a licence on WBF, did you find it ?",
          "votes": 2
        },
        {
          "id": 929051,
          "postDate": "2020-07-14T12:27:12.067Z",
          "content": "<p>here is WBF LICENSE. <br>\n<a href=\"https://github.com/ZFTurbo/Weighted-Boxes-Fusion/blob/master/LICENSE\">https://github.com/ZFTurbo/Weighted-Boxes-Fusion/blob/master/LICENSE</a></p>",
          "rawMarkdown": "here is WBF LICENSE.  \nhttps://github.com/ZFTurbo/Weighted-Boxes-Fusion/blob/master/LICENSE",
          "votes": 5
        },
        {
          "id": 929061,
          "postDate": "2020-07-14T12:46:06.363Z",
          "content": "<p>Oh nice! They added a license :)</p>",
          "rawMarkdown": "Oh nice! They added a license :)",
          "votes": 4
        },
        {
          "id": 930960,
          "postDate": "2020-07-15T20:57:23.633Z",
          "content": "<p><a href=\"/matthewmasters\">@matthewmasters</a> weird place to put this message I know, but I tried dropping you a PM via Kaggle's contact option but not sure if it went through. Let me know how I can contact you if not</p>",
          "rawMarkdown": "@matthewmasters weird place to put this message I know, but I tried dropping you a PM via Kaggle's contact option but not sure if it went through. Let me know how I can contact you if not",
          "votes": 1
        },
        {
          "id": 930962,
          "postDate": "2020-07-15T21:02:32.980Z",
          "content": "<p><a href=\"/alexandersoare\">@alexandersoare</a> Hey, I did get it! I'm sorry for not responding sooner. I will reply to you through email now</p>",
          "rawMarkdown": "@alexandersoare Hey, I did get it! I'm sorry for not responding sooner. I will reply to you through email now"
        }
      ]
    },
    {
      "id": 928530,
      "postDate": "2020-07-14T04:19:14.413Z",
      "content": "<p>ours .7501</p>",
      "rawMarkdown": "ours .7501",
      "replies": [
        {
          "id": 928533,
          "postDate": "2020-07-14T04:22:09.877Z",
          "content": "<p>Great, I'll update it now. I will put the score rather than the current ranking, that makes more sense.</p>",
          "rawMarkdown": "Great, I'll update it now. I will put the score rather than the current ranking, that makes more sense."
        }
      ]
    },
    {
      "id": 958145,
      "postDate": "2020-08-04T19:53:06.963Z",
      "content": "<p>I know I'm in a bit late here but I'm mostly just learning about EfficientDet. I have managed to score a 0.7362 in my meanderings. Looks like I'm not in too bad of shape compared to the rest of you.</p>\n\n<p>Good luck in the private board!</p>",
      "rawMarkdown": "I know I'm in a bit late here but I'm mostly just learning about EfficientDet. I have managed to score a 0.7362 in my meanderings. Looks like I'm not in too bad of shape compared to the rest of you.\n\nGood luck in the private board!"
    },
    {
      "id": 955252,
      "postDate": "2020-08-02T13:44:55.587Z",
      "content": "<p>0.7591 using EfficientDet5 + PL + TTA</p>",
      "rawMarkdown": "0.7591 using EfficientDet5 + PL + TTA",
      "replies": [
        {
          "id": 955294,
          "postDate": "2020-08-02T14:31:26.683Z",
          "content": "<p>Hi,Did you train long time on PL ,such as 7 epochs?</p>",
          "rawMarkdown": "Hi,Did you train long time on PL ,such as 7 epochs?"
        }
      ]
    },
    {
      "id": 954636,
      "postDate": "2020-08-02T00:10:27.953Z",
      "content": "<p>HI,have you added all scores to your board?</p>",
      "rawMarkdown": "HI,have you added all scores to your board?",
      "replies": [
        {
          "id": 954709,
          "postDate": "2020-08-02T02:38:44.897Z",
          "content": "<p>No, I will do one more update likely tomorrow.</p>",
          "rawMarkdown": "No, I will do one more update likely tomorrow."
        }
      ]
    },
    {
      "id": 953681,
      "postDate": "2020-08-01T02:22:09.373Z",
      "content": "<p>[Update] EfficientDet 0.7552, alan mise team's compliant entry. Thanks a lot!</p>",
      "rawMarkdown": "[Update] EfficientDet 0.7552, alan mise team's compliant entry. Thanks a lot!"
    },
    {
      "id": 953473,
      "postDate": "2020-07-31T20:11:59.350Z",
      "content": "<p>0.7434 with EfficientDet5</p>",
      "rawMarkdown": "0.7434 with EfficientDet5"
    },
    {
      "id": 940567,
      "postDate": "2020-07-23T03:04:48.670Z",
      "content": "<p>effdet 0.7398</p>",
      "rawMarkdown": "effdet 0.7398"
    },
    {
      "id": 939141,
      "postDate": "2020-07-22T04:06:03.993Z",
      "content": "<p>EfficientDet allowed? i remember effdet's license is gpl3.0👀  <a href=\"/matthewmasters\">@matthewmasters</a> </p>",
      "rawMarkdown": "EfficientDet allowed? i remember effdet's license is gpl3.0👀  @matthewmasters ",
      "replies": [
        {
          "id": 939187,
          "postDate": "2020-07-22T04:44:57.827Z",
          "content": "<p>The pytorch implementation that everyone is using is Apache 2.0 which is compatible with the MIT requirement</p>\n\n<p>Edit: had to change pyrotechnics to pytorch 😂 dang you autocorrect</p>",
          "rawMarkdown": "The pytorch implementation that everyone is using is Apache 2.0 which is compatible with the MIT requirement\n\nEdit: had to change pyrotechnics to pytorch 😂 dang you autocorrect"
        },
        {
          "id": 939258,
          "postDate": "2020-07-22T05:57:14.310Z",
          "content": "<p><a href=\"/matthewmasters\">@matthewmasters</a> Thank you for reply, what's the meaning of pyrotechnics?😲 </p>",
          "rawMarkdown": "@matthewmasters Thank you for reply, what's the meaning of pyrotechnics?😲 "
        },
        {
          "id": 939279,
          "postDate": "2020-07-22T06:20:33.177Z",
          "content": "<p>Haha it is a word that means fireworks or explosions. I typed my reply on my phone and it was autocorrected to that </p>",
          "rawMarkdown": "Haha it is a word that means fireworks or explosions. I typed my reply on my phone and it was autocorrected to that "
        },
        {
          "id": 939590,
          "postDate": "2020-07-22T10:15:09.127Z",
          "content": "<p>Sorry, I’m still confused. So Does effdet <a href=\"https://github.com/rwightman/pytorch-image-models\">rwightman's version which license is Apache 2.0</a> allowed? <a href=\"/matthewmasters\">@matthewmasters</a> </p>",
          "rawMarkdown": "Sorry, I’m still confused. So Does effdet [rwightman's version which license is Apache 2.0](https://github.com/rwightman/pytorch-image-models) allowed? @matthewmasters "
        },
        {
          "id": 939813,
          "postDate": "2020-07-22T13:31:28.930Z",
          "content": "<p>Yeah. Like I said, Apache 2.0 is allowed. Look at this post by Ross to clear up your confusion: <a href=\"https://www.kaggle.com/c/global-wheat-detection/discussion/164845#923524\">https://www.kaggle.com/c/global-wheat-detection/discussion/164845#923524</a></p>",
          "rawMarkdown": "Yeah. Like I said, Apache 2.0 is allowed. Look at this post by Ross to clear up your confusion: https://www.kaggle.com/c/global-wheat-detection/discussion/164845#923524",
          "votes": 1
        }
      ]
    },
    {
      "id": 937846,
      "postDate": "2020-07-21T07:37:37.123Z",
      "content": "<p>efficientdet 0.7379</p>",
      "rawMarkdown": "efficientdet 0.7379"
    },
    {
      "id": 934967,
      "postDate": "2020-07-19T01:00:53.247Z",
      "rawMarkdown": "",
      "isDeleted": true
    }
  ],
  "comments": [
    {
      "id": 936089,
      "author_name": "DungNB",
      "author_url": "",
      "post_date": "2020-07-20T01:00:53.117000",
      "content": "<p>0.7690 EfficientDet</p>",
      "votes": 11,
      "replies": [
        {
          "id": 939955,
          "author_name": "Marcin Czelej",
          "author_url": "",
          "post_date": "2020-07-22T15:33:10.300000",
          "content": "<p>Did you get this with Pseudolabeling or without ?</p>",
          "votes": 0,
          "replies": []
        },
        {
          "id": 939961,
          "author_name": "DungNB",
          "author_url": "",
          "post_date": "2020-07-22T15:41:18.340000",
          "content": "<p>with pseudolabeling</p>",
          "votes": 0,
          "replies": []
        },
        {
          "id": 943125,
          "author_name": "Dapeng",
          "author_url": "",
          "post_date": "2020-07-24T06:48:27.613000",
          "content": "<p><a href=\"/nguyenbadung\">@nguyenbadung</a>  Could u tell me the score without pseudo labeling?😄 Thanks</p>",
          "votes": 0,
          "replies": []
        },
        {
          "id": 943153,
          "author_name": "",
          "author_url": "",
          "post_date": "2020-07-24T07:04:04.757000",
          "content": "",
          "votes": 0,
          "replies": []
        },
        {
          "id": 943162,
          "author_name": "",
          "author_url": "",
          "post_date": "2020-07-24T07:12:49.880000",
          "content": "",
          "votes": 0,
          "replies": []
        },
        {
          "id": 947042,
          "author_name": "Peiyuan Liao",
          "author_url": "",
          "post_date": "2020-07-27T03:28:23.483000",
          "content": "<p>My EfficientDet based solution is at 0.765 without pseudo-labeling.</p>",
          "votes": 8,
          "replies": []
        },
        {
          "id": 951739,
          "author_name": "DungNB",
          "author_url": "",
          "post_date": "2020-07-30T11:03:42.573000",
          "content": "<p>Dapeng Lee, 0.7629 efficientdet without pseudolabeling</p>",
          "votes": 1,
          "replies": []
        },
        {
          "id": 951853,
          "author_name": "Dapeng",
          "author_url": "",
          "post_date": "2020-07-30T13:01:35.187000",
          "content": "<p><a href=\"/nguyenbadung\">@nguyenbadung</a> , <a href=\"/alexanderliao\">@alexanderliao</a>  Thanks!😄 our team's efficientdet have no improve with pseudolabeling ,maybe you guys find some tricks? haha</p>",
          "votes": 0,
          "replies": []
        }
      ]
    },
    {
      "id": 958123,
      "author_name": "Matt",
      "author_url": "",
      "post_date": "2020-08-04T19:39:04.073000",
      "content": "<p>Alright everyone, I've made my final update to the compliant leaderboard (hopefully there's no errors)! It's certainly grown a lot bigger than I initially thought it would be so I appreciate everyone who contributed their score. This competition has been an interesting  learning experience for many of us. Thank you to everyone who made this competition fun and enjoyable despite the setbacks and issues with licensing. And best of luck to all on the final leaderboard standings!</p>",
      "votes": 10,
      "replies": [
        {
          "id": 958146,
          "author_name": "Stanley Zheng",
          "author_url": "",
          "post_date": "2020-08-04T19:54:01.403000",
          "content": "<p>Thank you so much! It's been quite the journey :) This private LB should be interesting. </p>",
          "votes": 1,
          "replies": []
        },
        {
          "id": 958157,
          "author_name": "TheGruffalo",
          "author_url": "",
          "post_date": "2020-08-04T19:59:58.590000",
          "content": "<p>Thanks for this. I've been following along even though I haven't submitted much. Nice to know where I stand amongst everyone listed here</p>",
          "votes": 2,
          "replies": []
        },
        {
          "id": 958175,
          "author_name": "MichaelP",
          "author_url": "",
          "post_date": "2020-08-04T20:16:10.300000",
          "content": "<p>Thanks Matthew for the original list and for all the painstaking updates!</p>",
          "votes": 3,
          "replies": []
        },
        {
          "id": 958185,
          "author_name": "Vishnu Subramanian",
          "author_url": "",
          "post_date": "2020-08-04T20:29:49.833000",
          "content": "<p>Thanks Matthew for putting in a lot of effort. </p>",
          "votes": 3,
          "replies": []
        },
        {
          "id": 958263,
          "author_name": "wasupandceacar",
          "author_url": "",
          "post_date": "2020-08-04T22:06:45.987000",
          "content": "<p>👍 </p>",
          "votes": 0,
          "replies": []
        }
      ]
    },
    {
      "id": 957287,
      "author_name": "Peiyuan Liao",
      "author_url": "",
      "post_date": "2020-08-04T07:45:40.457000",
      "content": "<p>Possibly my last submission: 0.7733 with EfficientDet-based model.</p>\n\n<p>Honestly even without all the YoloV5 subs there's still gonna be a huge shakeup. The exact settings will give out a score that varies within the 0.003 range..... Which translates to currently a gold medal and something ranking 60+.</p>",
      "votes": 7,
      "replies": [
        {
          "id": 957309,
          "author_name": "bendang",
          "author_url": "",
          "post_date": "2020-08-04T08:01:57.707000",
          "content": "<p>Is this using a single model?</p>",
          "votes": 0,
          "replies": []
        }
      ]
    },
    {
      "id": 957959,
      "author_name": "Miras Amir",
      "author_url": "",
      "post_date": "2020-08-04T16:47:39.813000",
      "content": "<p>Our mmdet-based score: 0.7736</p>",
      "votes": 8,
      "replies": []
    },
    {
      "id": 953259,
      "author_name": "Neuron Engineer",
      "author_url": "",
      "post_date": "2020-07-31T16:25:47.963000",
      "content": "<p>Before deadline : 7638 [latest updated 7662] w/ FRCNN ResNest :) \n<strong>Edit</strong> answer <a href=\"/nvnnghia\">@nvnnghia</a> below, it's with pseudo labeling (thanks to your contributions). <a href=\"/gengzl\">@gengzl</a> As deadline approaces, we cannot share details, just the main idea that we use the best public ResNest, followed by <a href=\"/nvnnghia\">@nvnnghia</a> pseudo labeling framework with some adaptations to the dataset </p>",
      "votes": 5,
      "replies": [
        {
          "id": 953315,
          "author_name": "Nitin Datta",
          "author_url": "",
          "post_date": "2020-07-31T17:22:41.327000",
          "content": "<p>In addition to this I think <a href=\"/kyoshioka47\">@kyoshioka47</a>  score can be removed....\nWe merged..\nNot to cause any confusion 😅 </p>",
          "votes": 2,
          "replies": []
        },
        {
          "id": 953596,
          "author_name": "nvnn",
          "author_url": "",
          "post_date": "2020-07-31T22:49:09.283000",
          "content": "<p>very impressive!! Is that with or without pseudo labeling</p>",
          "votes": 2,
          "replies": []
        },
        {
          "id": 953672,
          "author_name": "",
          "author_url": "",
          "post_date": "2020-08-01T01:37:24.983000",
          "content": "",
          "votes": 0,
          "replies": []
        }
      ]
    },
    {
      "id": 941469,
      "author_name": "tereka",
      "author_url": "",
      "post_date": "2020-07-23T08:27:39.053000",
      "content": "<p>0.7700.(Effdet)</p>",
      "votes": 5,
      "replies": [
        {
          "id": 944903,
          "author_name": "tereka",
          "author_url": "",
          "post_date": "2020-07-25T12:45:35.007000",
          "content": "<p>Update: 0.7730(Effdet)</p>",
          "votes": 5,
          "replies": []
        },
        {
          "id": 947935,
          "author_name": "Manh Lab",
          "author_url": "",
          "post_date": "2020-07-27T15:03:29.447000",
          "content": "<p>which version effdet you used?</p>",
          "votes": 0,
          "replies": []
        },
        {
          "id": 953588,
          "author_name": "tereka",
          "author_url": "",
          "post_date": "2020-07-31T22:36:11.097000",
          "content": "<p>Update: 0.7753(Effdet)</p>\n\n<p>&gt; which version effdet you used?  </p>\n\n<p>It's secret until over deadline.  </p>",
          "votes": 3,
          "replies": []
        }
      ]
    },
    {
      "id": 953614,
      "author_name": "Vishnu Subramanian",
      "author_url": "",
      "post_date": "2020-07-31T23:16:22.477000",
      "content": "<p>0.7514 - Retinanet - With EfficientNet_backbone :)</p>",
      "votes": 3,
      "replies": []
    },
    {
      "id": 949647,
      "author_name": "Darkate",
      "author_url": "",
      "post_date": "2020-07-28T19:09:45.867000",
      "content": "<p>0.7484 CenterNet, 8:2 single split, TTAx3 (fliplr &amp; flipud), no pseudo label\n[Update] 0.7520 with TTAx9\n[Update 2] 0.7588 with 3 folds &amp;, TTAx6\n[Final Update] 0.7590 with 3 folds, TTAx6; no luck with pseudo label ☹️ </p>",
      "votes": 3,
      "replies": [
        {
          "id": 951624,
          "author_name": "Libo Wang",
          "author_url": "",
          "post_date": "2020-07-30T09:15:06.243000",
          "content": "<p>what's the meaning of TTAx9?</p>",
          "votes": 0,
          "replies": []
        },
        {
          "id": 952026,
          "author_name": "ProtsenkoAI",
          "author_url": "",
          "post_date": "2020-07-30T15:08:53.103000",
          "content": "<p>TTA is test time augmentation, I think 9 is 9 augmentations per photo (but maybe not)</p>",
          "votes": 0,
          "replies": []
        },
        {
          "id": 953503,
          "author_name": "Darkate",
          "author_url": "",
          "post_date": "2020-07-31T20:54:55.897000",
          "content": "<p>Original, flip, multi-scale. 8 augmentations per image + original</p>",
          "votes": 0,
          "replies": []
        },
        {
          "id": 953627,
          "author_name": "Libo Wang",
          "author_url": "",
          "post_date": "2020-07-31T23:38:52.290000",
          "content": "<p>Thank you for your sharing👍 </p>",
          "votes": 0,
          "replies": []
        }
      ]
    },
    {
      "id": 947926,
      "author_name": "Rauf  Yagfarov",
      "author_url": "",
      "post_date": "2020-07-27T15:00:20.507000",
      "content": "<p>0.7573 mmdetection models \n[Update] 0.7601</p>",
      "votes": 3,
      "replies": []
    },
    {
      "id": 930433,
      "author_name": "Anton Sakharov",
      "author_url": "",
      "post_date": "2020-07-15T13:06:40.203000",
      "content": "<p>I have fastercnn with 0.7491</p>",
      "votes": 3,
      "replies": [
        {
          "id": 930859,
          "author_name": "Matt",
          "author_url": "",
          "post_date": "2020-07-15T18:52:36.067000",
          "content": "<p>Added!</p>",
          "votes": 0,
          "replies": []
        }
      ]
    },
    {
      "id": 929903,
      "author_name": "wasupandceacar",
      "author_url": "",
      "post_date": "2020-07-15T04:11:13.183000",
      "content": "<p>0.7601 EfficientDet</p>",
      "votes": 3,
      "replies": [
        {
          "id": 929923,
          "author_name": "Matt",
          "author_url": "",
          "post_date": "2020-07-15T04:46:15.260000",
          "content": "<p>Thanks, added to the list!</p>",
          "votes": 0,
          "replies": []
        }
      ]
    },
    {
      "id": 957130,
      "author_name": "Nacir Bouazizi",
      "author_url": "",
      "post_date": "2020-08-04T04:48:59.937000",
      "content": "<p>0.7560 EffdetD5 </p>",
      "votes": 1,
      "replies": []
    },
    {
      "id": 957091,
      "author_name": "Ciel 42",
      "author_url": "",
      "post_date": "2020-08-04T03:39:08.547000",
      "content": "<p>0.7597 with effdet d5</p>",
      "votes": 1,
      "replies": []
    },
    {
      "id": 956477,
      "author_name": "( ͡° ͜ʖ ͡°)",
      "author_url": "",
      "post_date": "2020-08-03T14:32:32.167000",
      "content": "<p>I'm compliant at 0.7438</p>",
      "votes": 1,
      "replies": []
    },
    {
      "id": 954849,
      "author_name": "Vee",
      "author_url": "",
      "post_date": "2020-08-02T06:21:45.473000",
      "content": "<p>[UPDATE] 0.7602 using EfficientDet5 backbone.</p>",
      "votes": 1,
      "replies": [
        {
          "id": 954853,
          "author_name": "Nitin Datta",
          "author_url": "",
          "post_date": "2020-08-02T06:28:23.037000",
          "content": "<p>Is this with pseudo-labeling??</p>",
          "votes": 0,
          "replies": []
        },
        {
          "id": 954856,
          "author_name": "Vee",
          "author_url": "",
          "post_date": "2020-08-02T06:32:50.440000",
          "content": "<p>Yes, with Pseduo-Labelling.</p>",
          "votes": 1,
          "replies": []
        },
        {
          "id": 954870,
          "author_name": "gao-hongnan",
          "author_url": "",
          "post_date": "2020-08-02T06:48:40",
          "content": "<p><a href=\"/watzisname\">@watzisname</a> oh brother, so this means your current best score is yolov5 but managed to get 0.7575 on effdet?</p>",
          "votes": 0,
          "replies": []
        },
        {
          "id": 955128,
          "author_name": "Vee",
          "author_url": "",
          "post_date": "2020-08-02T11:22:32.547000",
          "content": "<p><a href=\"/reighns\">@reighns</a>\nYes that is correct.</p>",
          "votes": 0,
          "replies": []
        }
      ]
    },
    {
      "id": 951192,
      "author_name": "Stanley Zheng",
      "author_url": "",
      "post_date": "2020-07-30T01:24:04.260000",
      "content": "<p>0.7496 effdet, 0.7115 yolov4 without PL or OOF. Thanks a ton Matthew!</p>",
      "votes": 1,
      "replies": []
    },
    {
      "id": 950889,
      "author_name": "Manoj Prabhakar",
      "author_url": "",
      "post_date": "2020-07-29T17:34:13.497000",
      "content": "<p><a href=\"/matthewmasters\">@matthewmasters</a> - I have been following your posts in this competition. You have been doing an amazing work. 0.7378 with EfficientDet\nUpdate: 0.7384 EfficientDet</p>",
      "votes": 1,
      "replies": [
        {
          "id": 951049,
          "author_name": "Matt",
          "author_url": "",
          "post_date": "2020-07-29T20:48:30.450000",
          "content": "<p>Thanks for the kind words. I added your score to the LB</p>",
          "votes": 1,
          "replies": []
        }
      ]
    },
    {
      "id": 950264,
      "author_name": "GaoLijing",
      "author_url": "",
      "post_date": "2020-07-29T09:57:49.770000",
      "content": "<p>0.7450 with EfficientDet</p>",
      "votes": 1,
      "replies": []
    },
    {
      "id": 945222,
      "author_name": "Matt",
      "author_url": "",
      "post_date": "2020-07-25T16:57:44.033000",
      "content": "<p>Sorry for the delay in updates. I have just updated the list again</p>",
      "votes": 1,
      "replies": []
    },
    {
      "id": 935574,
      "author_name": "Tiago de Miranda Leite",
      "author_url": "",
      "post_date": "2020-07-19T13:55:50.440000",
      "content": "<p>YoloV3 0.7509, my compliant entry.</p>",
      "votes": 1,
      "replies": [
        {
          "id": 951183,
          "author_name": "",
          "author_url": "",
          "post_date": "2020-07-30T01:15:36.083000",
          "content": "",
          "votes": 0,
          "replies": []
        }
      ]
    },
    {
      "id": 934969,
      "author_name": "Matt",
      "author_url": "",
      "post_date": "2020-07-19T01:01:50.260000",
      "content": "<p><a href=\"/nvnnghia\">@nvnnghia</a> Very nice work! I updated the listing</p>\n\n<p>(not sure why but it wouldn't let me reply directly to your comment)</p>",
      "votes": 1,
      "replies": []
    },
    {
      "id": 934958,
      "author_name": "nvnn",
      "author_url": "",
      "post_date": "2020-07-19T00:34:26.610000",
      "content": "<p>[UPDATED] 0.7683</p>",
      "votes": 1,
      "replies": []
    },
    {
      "id": 932400,
      "author_name": "alan chen",
      "author_url": "",
      "post_date": "2020-07-17T03:29:22.220000",
      "content": "<p>EfficientDet (no psuedo) 0.7428, alan mise team's compliant entry.</p>",
      "votes": 1,
      "replies": [
        {
          "id": 932679,
          "author_name": "Matt",
          "author_url": "",
          "post_date": "2020-07-17T07:46:50.560000",
          "content": "<p>Forgot to respond but I added this earlier! :)</p>",
          "votes": 2,
          "replies": []
        }
      ]
    },
    {
      "id": 932362,
      "author_name": "mrlcv",
      "author_url": "",
      "post_date": "2020-07-17T02:31:31.220000",
      "content": "<p>EfficientDet: 0.7365</p>",
      "votes": 1,
      "replies": [
        {
          "id": 932364,
          "author_name": "Matt",
          "author_url": "",
          "post_date": "2020-07-17T02:34:54.987000",
          "content": "<p>Added</p>",
          "votes": 1,
          "replies": []
        }
      ]
    },
    {
      "id": 931935,
      "author_name": "ZaoDianNao",
      "author_url": "",
      "post_date": "2020-07-16T15:06:37.520000",
      "content": "<p>EfficientDet: 0.7382; My compliant entry.</p>",
      "votes": 1,
      "replies": [
        {
          "id": 932099,
          "author_name": "Matt",
          "author_url": "",
          "post_date": "2020-07-16T17:37:54.747000",
          "content": "<p>Added!</p>",
          "votes": 0,
          "replies": []
        }
      ]
    },
    {
      "id": 931759,
      "author_name": "gzYou",
      "author_url": "",
      "post_date": "2020-07-16T12:30:52.310000",
      "content": "<p>0.7510 EfficientDet\n[update]0.7582</p>",
      "votes": 1,
      "replies": [
        {
          "id": 932100,
          "author_name": "Matt",
          "author_url": "",
          "post_date": "2020-07-16T17:38:02.170000",
          "content": "<p>Added!</p>",
          "votes": 0,
          "replies": []
        },
        {
          "id": 932674,
          "author_name": "gzYou",
          "author_url": "",
          "post_date": "2020-07-17T07:40:19.547000",
          "content": "<p>Thanks <a href=\"/matthewmasters\">@matthewmasters</a> . Maybe there should be a positive contribution award for people like you who contribute to the community</p>",
          "votes": 3,
          "replies": []
        },
        {
          "id": 932678,
          "author_name": "Matt",
          "author_url": "",
          "post_date": "2020-07-17T07:46:26.980000",
          "content": "<p>Well I'm glad you appreciate my contribution, that means a lot! Actually this morning I got an email from a member of the Kaggle team to send me swag for my active participation, so that's was certainly nice to receive! 😀 </p>",
          "votes": 2,
          "replies": []
        },
        {
          "id": 932685,
          "author_name": "Mr_KnowNothing",
          "author_url": "",
          "post_date": "2020-07-17T07:49:08.863000",
          "content": "<p><a href=\"/guizengyou\">@guizengyou</a> Totally agreed \n<a href=\"/matthewmasters\">@matthewmasters</a>  I am really glad to hear that , you deserve it bro</p>",
          "votes": 1,
          "replies": []
        },
        {
          "id": 947893,
          "author_name": "Manh Lab",
          "author_url": "",
          "post_date": "2020-07-27T14:41:16.877000",
          "content": "<p>Did you use D5 or D7</p>",
          "votes": 0,
          "replies": []
        }
      ]
    },
    {
      "id": 928816,
      "author_name": "Marcin Czelej",
      "author_url": "",
      "post_date": "2020-07-14T08:35:10.700000",
      "content": "<p>EfficientDet ( I think it`s ok ) 0.7548</p>\n\n<p>Update: 0.7614 without PseudoLabeling</p>",
      "votes": 1,
      "replies": [
        {
          "id": 928820,
          "author_name": "Matt",
          "author_url": "",
          "post_date": "2020-07-14T08:37:42.520000",
          "content": "<p>Thanks, added! PS. EfficientDet is okay but has some specific requirements. The creator of the pytorch EfficientDet implementation can be found <a href=\"https://www.kaggle.com/c/global-wheat-detection/discussion/164845#923524\">here</a> giving some advice to this end.</p>",
          "votes": 1,
          "replies": []
        },
        {
          "id": 928823,
          "author_name": "Marcin Czelej",
          "author_url": "",
          "post_date": "2020-07-14T08:41:00.390000",
          "content": "<p>Thanks :) Good to know I`ll add this.</p>\n\n<p>BTW, can someone please write what models are ok ? Because right now it`s hard to catch up with all the informations</p>",
          "votes": 1,
          "replies": []
        },
        {
          "id": 928824,
          "author_name": "Matt",
          "author_url": "",
          "post_date": "2020-07-14T08:44:38.463000",
          "content": "<p>Yes actually! <a href=\"/nakajima\">@nakajima</a> put together <a href=\"https://www.kaggle.com/c/global-wheat-detection/discussion/165735#924609\">this list</a> which should help you with that</p>",
          "votes": 0,
          "replies": []
        }
      ]
    },
    {
      "id": 928804,
      "author_name": "Chanhu",
      "author_url": "",
      "post_date": "2020-07-14T08:20:42.450000",
      "content": "<p>I'm compliant at 0.7669</p>",
      "votes": 1,
      "replies": [
        {
          "id": 928805,
          "author_name": "Matt",
          "author_url": "",
          "post_date": "2020-07-14T08:22:44.030000",
          "content": "<p>Well done! I added it to the list</p>",
          "votes": 0,
          "replies": []
        }
      ]
    },
    {
      "id": 928733,
      "author_name": "Ryan Epp",
      "author_url": "",
      "post_date": "2020-07-14T07:18:07.287000",
      "content": "<p>I'm compliant at 0.7529</p>",
      "votes": 1,
      "replies": [
        {
          "id": 928739,
          "author_name": "Matt",
          "author_url": "",
          "post_date": "2020-07-14T07:20:46.517000",
          "content": "<p>Good work! I just added you to the list :)</p>",
          "votes": 1,
          "replies": []
        }
      ]
    },
    {
      "id": 956777,
      "author_name": "",
      "author_url": "",
      "post_date": "2020-08-03T19:10:07.823000",
      "content": "",
      "votes": 2,
      "replies": []
    },
    {
      "id": 954039,
      "author_name": "",
      "author_url": "",
      "post_date": "2020-08-01T10:30:02.237000",
      "content": "",
      "votes": 2,
      "replies": []
    },
    {
      "id": 951048,
      "author_name": "",
      "author_url": "",
      "post_date": "2020-07-29T20:48:09.847000",
      "content": "",
      "votes": 2,
      "replies": [
        {
          "id": 953429,
          "author_name": "",
          "author_url": "",
          "post_date": "2020-07-31T19:32:35.747000",
          "content": "",
          "votes": 4,
          "replies": []
        }
      ]
    },
    {
      "id": 946142,
      "author_name": "",
      "author_url": "",
      "post_date": "2020-07-26T11:38:56.287000",
      "content": "",
      "votes": 2,
      "replies": []
    },
    {
      "id": 946031,
      "author_name": "",
      "author_url": "",
      "post_date": "2020-07-26T09:51:33.177000",
      "content": "",
      "votes": 2,
      "replies": []
    },
    {
      "id": 940717,
      "author_name": "",
      "author_url": "",
      "post_date": "2020-07-23T05:24:41.560000",
      "content": "",
      "votes": 2,
      "replies": []
    },
    {
      "id": 936377,
      "author_name": "",
      "author_url": "",
      "post_date": "2020-07-20T07:12:20.190000",
      "content": "",
      "votes": 2,
      "replies": []
    },
    {
      "id": 936369,
      "author_name": "",
      "author_url": "",
      "post_date": "2020-07-20T07:06:40.510000",
      "content": "",
      "votes": 2,
      "replies": []
    },
    {
      "id": 929419,
      "author_name": "",
      "author_url": "",
      "post_date": "2020-07-14T16:52:29.737000",
      "content": "",
      "votes": 2,
      "replies": [
        {
          "id": 929621,
          "author_name": "",
          "author_url": "",
          "post_date": "2020-07-14T19:52:12.893000",
          "content": "",
          "votes": 2,
          "replies": []
        },
        {
          "id": 929657,
          "author_name": "",
          "author_url": "",
          "post_date": "2020-07-14T20:27:45.413000",
          "content": "",
          "votes": 1,
          "replies": []
        },
        {
          "id": 929679,
          "author_name": "",
          "author_url": "",
          "post_date": "2020-07-14T20:58:45.950000",
          "content": "",
          "votes": 8,
          "replies": []
        },
        {
          "id": 929696,
          "author_name": "",
          "author_url": "",
          "post_date": "2020-07-14T21:25:59.300000",
          "content": "",
          "votes": 8,
          "replies": []
        },
        {
          "id": 929773,
          "author_name": "",
          "author_url": "",
          "post_date": "2020-07-14T23:48:52.147000",
          "content": "",
          "votes": 0,
          "replies": []
        },
        {
          "id": 929909,
          "author_name": "",
          "author_url": "",
          "post_date": "2020-07-15T04:17:44.970000",
          "content": "",
          "votes": 1,
          "replies": []
        },
        {
          "id": 930270,
          "author_name": "",
          "author_url": "",
          "post_date": "2020-07-15T10:10:28.967000",
          "content": "",
          "votes": 4,
          "replies": []
        },
        {
          "id": 930302,
          "author_name": "",
          "author_url": "",
          "post_date": "2020-07-15T10:49:18.480000",
          "content": "",
          "votes": 3,
          "replies": []
        },
        {
          "id": 930862,
          "author_name": "",
          "author_url": "",
          "post_date": "2020-07-15T18:54:49.573000",
          "content": "",
          "votes": 1,
          "replies": []
        },
        {
          "id": 930895,
          "author_name": "",
          "author_url": "",
          "post_date": "2020-07-15T19:45:52.700000",
          "content": "",
          "votes": 1,
          "replies": []
        },
        {
          "id": 945574,
          "author_name": "",
          "author_url": "",
          "post_date": "2020-07-26T01:16:27.363000",
          "content": "",
          "votes": 2,
          "replies": []
        }
      ]
    },
    {
      "id": 928828,
      "author_name": "",
      "author_url": "",
      "post_date": "2020-07-14T08:47:15.883000",
      "content": "",
      "votes": 2,
      "replies": [
        {
          "id": 928834,
          "author_name": "",
          "author_url": "",
          "post_date": "2020-07-14T08:50:42.963000",
          "content": "",
          "votes": 0,
          "replies": []
        }
      ]
    },
    {
      "id": 928716,
      "author_name": "",
      "author_url": "",
      "post_date": "2020-07-14T07:10:41.380000",
      "content": "",
      "votes": 2,
      "replies": [
        {
          "id": 928722,
          "author_name": "",
          "author_url": "",
          "post_date": "2020-07-14T07:13:56.860000",
          "content": "",
          "votes": 0,
          "replies": []
        },
        {
          "id": 929017,
          "author_name": "",
          "author_url": "",
          "post_date": "2020-07-14T11:59:23.270000",
          "content": "",
          "votes": 2,
          "replies": []
        },
        {
          "id": 929051,
          "author_name": "",
          "author_url": "",
          "post_date": "2020-07-14T12:27:12.067000",
          "content": "",
          "votes": 5,
          "replies": []
        },
        {
          "id": 929061,
          "author_name": "",
          "author_url": "",
          "post_date": "2020-07-14T12:46:06.363000",
          "content": "",
          "votes": 4,
          "replies": []
        },
        {
          "id": 930960,
          "author_name": "",
          "author_url": "",
          "post_date": "2020-07-15T20:57:23.633000",
          "content": "",
          "votes": 1,
          "replies": []
        },
        {
          "id": 930962,
          "author_name": "",
          "author_url": "",
          "post_date": "2020-07-15T21:02:32.980000",
          "content": "",
          "votes": 0,
          "replies": []
        }
      ]
    },
    {
      "id": 928530,
      "author_name": "",
      "author_url": "",
      "post_date": "2020-07-14T04:19:14.413000",
      "content": "",
      "votes": 0,
      "replies": [
        {
          "id": 928533,
          "author_name": "",
          "author_url": "",
          "post_date": "2020-07-14T04:22:09.877000",
          "content": "",
          "votes": 0,
          "replies": []
        }
      ]
    },
    {
      "id": 958145,
      "author_name": "",
      "author_url": "",
      "post_date": "2020-08-04T19:53:06.963000",
      "content": "",
      "votes": 0,
      "replies": []
    },
    {
      "id": 955252,
      "author_name": "",
      "author_url": "",
      "post_date": "2020-08-02T13:44:55.587000",
      "content": "",
      "votes": 0,
      "replies": [
        {
          "id": 955294,
          "author_name": "",
          "author_url": "",
          "post_date": "2020-08-02T14:31:26.683000",
          "content": "",
          "votes": 0,
          "replies": []
        }
      ]
    },
    {
      "id": 954636,
      "author_name": "",
      "author_url": "",
      "post_date": "2020-08-02T00:10:27.953000",
      "content": "",
      "votes": 0,
      "replies": [
        {
          "id": 954709,
          "author_name": "",
          "author_url": "",
          "post_date": "2020-08-02T02:38:44.897000",
          "content": "",
          "votes": 0,
          "replies": []
        }
      ]
    },
    {
      "id": 953681,
      "author_name": "",
      "author_url": "",
      "post_date": "2020-08-01T02:22:09.373000",
      "content": "",
      "votes": 0,
      "replies": []
    },
    {
      "id": 953473,
      "author_name": "",
      "author_url": "",
      "post_date": "2020-07-31T20:11:59.350000",
      "content": "",
      "votes": 0,
      "replies": []
    },
    {
      "id": 940567,
      "author_name": "",
      "author_url": "",
      "post_date": "2020-07-23T03:04:48.670000",
      "content": "",
      "votes": 0,
      "replies": []
    },
    {
      "id": 939141,
      "author_name": "",
      "author_url": "",
      "post_date": "2020-07-22T04:06:03.993000",
      "content": "",
      "votes": 0,
      "replies": [
        {
          "id": 939187,
          "author_name": "",
          "author_url": "",
          "post_date": "2020-07-22T04:44:57.827000",
          "content": "",
          "votes": 0,
          "replies": []
        },
        {
          "id": 939258,
          "author_name": "",
          "author_url": "",
          "post_date": "2020-07-22T05:57:14.310000",
          "content": "",
          "votes": 0,
          "replies": []
        },
        {
          "id": 939279,
          "author_name": "",
          "author_url": "",
          "post_date": "2020-07-22T06:20:33.177000",
          "content": "",
          "votes": 0,
          "replies": []
        },
        {
          "id": 939590,
          "author_name": "",
          "author_url": "",
          "post_date": "2020-07-22T10:15:09.127000",
          "content": "",
          "votes": 0,
          "replies": []
        },
        {
          "id": 939813,
          "author_name": "",
          "author_url": "",
          "post_date": "2020-07-22T13:31:28.930000",
          "content": "",
          "votes": 1,
          "replies": []
        }
      ]
    },
    {
      "id": 937846,
      "author_name": "",
      "author_url": "",
      "post_date": "2020-07-21T07:37:37.123000",
      "content": "",
      "votes": 0,
      "replies": []
    },
    {
      "id": 934967,
      "author_name": "",
      "author_url": "",
      "post_date": "2020-07-19T01:00:53.247000",
      "content": "",
      "votes": 0,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "928485": "Right now it's very difficult to tell what the current leaderboard standings are due to [the licensing issue](https://www.kaggle.com/c/global-wheat-detection/discussion/165735). As @alexandersoare suggested ([thread](https://www.kaggle.com/c/global-wheat-detection/discussion/165424)), some competitors have changed their team name to indicate their compliance. However, it's still hard to tell as some people have compliant solutions but their highest LB score isn't. So, I would like to use this thread to gather the highest compliant LB scores. The current standings (to the best of my knowledge) are as follows:\n\n1. 0.7753 @tereka \n2. 0.7736 @amiras \n3. 0.7733 @alexanderliao \n4. 0.7695 @datahobbit and team\n5. 0.7690 @nguyenbadung \n6. 0.7680 @chanhu \n7. 0.7662 @ratthachat \n8. 0.7630 @nvnnghia and team\n9. 0.7625 @li325040229 \n10. 0.7602 @watzisname \n11. 0.7601 @wasupandceacar \n12. 0.7601 @raufyagfarov \n13. 0.7597 @ciel42 \n14. 0.7591 @creatrol \n15. 0.7582 @guizengyou \n16. 0.7560 @nacirbouazizi \n17. 0.7553 @phalanx \n18. 0.7552 @alanchen1115 \n19. 0.7550 @voglinio\n20. 0.7548 @marcinczelej \n21. 0.7541 @bendang \n22. 0.7536 @alexandersoare \n23. 0.7529 @reppic \n24. 0.7514 @vishnus \n25. 0.7509 @tiagotml \n26. 0.7496 @stanleyjzheng \n27. 0.7491 @sanantoha \n28. 0.7484 @jihangz \n30. 0.7479 @whurobin \n31. 0.7464 myself\n32. 0.7451 @ufownl \n33. 0.7450 @gaolijing \n34. 0.7438 @truonghoang \n35. 0.7434 @zhaonat \n36. 0.7428 @alanchen1115 and team\n37. 0.7398 @cswwp347724 \n38. 0.7384 @manojprabhaakr \n39. 0.7382 @thomasdic2000 \n40. 0.7379 @sistmrl \n\nThis is obviously an incomplete list and I have undoubtedly missed many of you. If you'd like to share your highest compliant LB score, just comment below and I will try to keep the list up to date!",
    "936089": "0.7690 EfficientDet",
    "958123": "Alright everyone, I've made my final update to the compliant leaderboard (hopefully there's no errors)! It's certainly grown a lot bigger than I initially thought it would be so I appreciate everyone who contributed their score. This competition has been an interesting  learning experience for many of us. Thank you to everyone who made this competition fun and enjoyable despite the setbacks and issues with licensing. And best of luck to all on the final leaderboard standings!",
    "957287": "Possibly my last submission: 0.7733 with EfficientDet-based model.\n\nHonestly even without all the YoloV5 subs there's still gonna be a huge shakeup. The exact settings will give out a score that varies within the 0.003 range..... Which translates to currently a gold medal and something ranking 60+.",
    "957959": "Our mmdet-based score: 0.7736",
    "953259": "Before deadline : 7638 [latest updated 7662] w/ FRCNN ResNest :) \n**Edit** answer @nvnnghia below, it's with pseudo labeling (thanks to your contributions). @gengzl As deadline approaces, we cannot share details, just the main idea that we use the best public ResNest, followed by @nvnnghia pseudo labeling framework with some adaptations to the dataset ",
    "941469": "0.7700.(Effdet)",
    "953614": "0.7514 - Retinanet - With EfficientNet_backbone :)",
    "949647": "0.7484 CenterNet, 8:2 single split, TTAx3 (fliplr &amp; flipud), no pseudo label\n[Update] 0.7520 with TTAx9\n[Update 2] 0.7588 with 3 folds &amp;, TTAx6\n[Final Update] 0.7590 with 3 folds, TTAx6; no luck with pseudo label ☹️ ",
    "947926": "0.7573 mmdetection models \n[Update] 0.7601",
    "930433": "I have fastercnn with 0.7491",
    "929903": "0.7601 EfficientDet",
    "957130": "0.7560 EffdetD5 ",
    "957091": "0.7597 with effdet d5",
    "956477": "I'm compliant at 0.7438",
    "954849": "[UPDATE] 0.7602 using EfficientDet5 backbone.",
    "951192": "0.7496 effdet, 0.7115 yolov4 without PL or OOF. Thanks a ton Matthew!",
    "950889": "@matthewmasters - I have been following your posts in this competition. You have been doing an amazing work. 0.7378 with EfficientDet\nUpdate: 0.7384 EfficientDet",
    "950264": "0.7450 with EfficientDet",
    "945222": "Sorry for the delay in updates. I have just updated the list again",
    "935574": "YoloV3 0.7509, my compliant entry.",
    "934969": "@nvnnghia Very nice work! I updated the listing\n\n(not sure why but it wouldn't let me reply directly to your comment)",
    "934958": "[UPDATED] 0.7683\n",
    "932400": "EfficientDet (no psuedo) 0.7428, alan mise team's compliant entry.",
    "932362": "EfficientDet: 0.7365",
    "931935": "EfficientDet: 0.7382; My compliant entry.",
    "931759": "0.7510 EfficientDet\n[update]0.7582",
    "928816": "EfficientDet ( I think it`s ok ) 0.7548\n\nUpdate: 0.7614 without PseudoLabeling",
    "928804": "I'm compliant at 0.7669",
    "928733": "I'm compliant at 0.7529",
    "956777": "[UPDATE]- 0.7695 FRCNN\nfrom @datahobbit and team...\nAlso @kyoshioka47's score can be removed as we are in same team...\nThank you",
    "954039": "@matthewmasters ,Thanks ,matthew! ours compliant entry: 0.7625.",
    "951048": "Updated again. Seeing as its the last week, I will likely only make one more update before the competition end",
    "946142": "0.7550 Effdet + TTAx16 + pseudo labelling",
    "946031": "0.7541 EffDet d5 + d7x + FRCNN\n[Update] 0.7591\n[Update] 0.7613",
    "940717": "effdet 0.7680....",
    "936377": "Updated with scores of @nguyenbadung, @whurobin, and @tiagotml ",
    "936369": "0.7466\n[UPDATED] 0.7479\n[UPDATED] 0.7551",
    "929419": "@matthewmasters you have been really active in this competition , I really thank you for all your discussion threads and work\nI would be glad if you clear my following doubts:-\n* If a team uses yolo will it be removed from the lb? or only the winning solutions should not use , but if you geta gold using yolo its ohk?  There is a lot of confusion around it\n* The images which the correction is done are present only in 38 percent hidden test set ? Also is there a dataset compiled by someone which contains all the images which have noisy bboxes?\n\nThanks in advance",
    "928828": "I'm compliant at 0.7613 :)",
    "928716": ".7536 mostly compliant apart from wbf. But I'm sure everyone has that problem and if they feel they are going to win it they'll just write their own implementation.",
    "928530": "ours .7501",
    "958145": "I know I'm in a bit late here but I'm mostly just learning about EfficientDet. I have managed to score a 0.7362 in my meanderings. Looks like I'm not in too bad of shape compared to the rest of you.\n\nGood luck in the private board!",
    "955252": "0.7591 using EfficientDet5 + PL + TTA",
    "954636": "HI,have you added all scores to your board?",
    "953681": "[Update] EfficientDet 0.7552, alan mise team's compliant entry. Thanks a lot!",
    "953473": "0.7434 with EfficientDet5",
    "940567": "effdet 0.7398",
    "939141": "EfficientDet allowed? i remember effdet's license is gpl3.0👀  @matthewmasters ",
    "937846": "efficientdet 0.7379",
    "934967": ""
  }
}