{
  "id": 58472,
  "title": "Thank LiBin Helps us to fix a new bug!",
  "url": "/competitions/cvpr-2018-autonomous-driving/discussion/58472",
  "author_name": "",
  "post_date": "2018-06-08T13:48:14.005534500Z",
  "votes": null,
  "comment_count": 31,
  "views": 0,
  "content": "<p>Thanks LiBin for Helping us fixed a new bug!</p>\n\n<p>&nbsp; &nbsp; &nbsp; I fixed a new bug. If user's&nbsp;PixelCount in&nbsp;submission is wrong, the result will be hacked.\n&nbsp; &nbsp; &nbsp;&nbsp;it because&nbsp;overlap = pred.intersection / (gt.pixelCount +&nbsp;pred.pixelCount&nbsp;- pred.intersection);\n&nbsp; &nbsp; &nbsp; It means if submission set&nbsp;all pixelCount = 0.&nbsp;&nbsp;the overlap will be much higher than it should be.&nbsp;</p>\n\n<p>&nbsp; &nbsp; &nbsp; I have changed code to let program count pixelCount by itself, instead of using user's.&nbsp;\n&nbsp; &nbsp; &nbsp; The function is&nbsp;&nbsp;DecodePridictionCVS_Kaggle(CVPRAutoDrivingAveragePrecision.Submission _submission)</p>\n\n<p>LiBin’s report:\nI used [\"ImageId\", \"LabelId\",\"PixelCount\", \"Confidence\", \"EncodedPixels\"] as csv’s header and the real data use this order([“ImageId\", \"LabelId\", \"Confidence\", \"PixelCount\", \"EncodedPixels”]), I could got a very high score(0.31),but when I use [“ImageId\", \"LabelId\", \"Confidence\", \"PixelCount\", \"EncodedPixels”] as header and real data’s order, I got a worse score(0.11).</p>",
  "messages": [
    {
      "id": "340159",
      "postDate": "06/08/2018 13:48:14",
      "content": "<p>Thanks LiBin for Helping us fixed a new bug!</p>\n\n<p>&nbsp; &nbsp; &nbsp; I fixed a new bug. If user's&nbsp;PixelCount in&nbsp;submission is wrong, the result will be hacked.\n&nbsp; &nbsp; &nbsp;&nbsp;it because&nbsp;overlap = pred.intersection / (gt.pixelCount +&nbsp;pred.pixelCount&nbsp;- pred.intersection);\n&nbsp; &nbsp; &nbsp; It means if submission set&nbsp;all pixelCount = 0.&nbsp;&nbsp;the overlap will be much higher than it should be.&nbsp;</p>\n\n<p>&nbsp; &nbsp; &nbsp; I have changed code to let program count pixelCount by itself, instead of using user's.&nbsp;\n&nbsp; &nbsp; &nbsp; The function is&nbsp;&nbsp;DecodePridictionCVS_Kaggle(CVPRAutoDrivingAveragePrecision.Submission _submission)</p>\n\n<p>LiBin’s report:\nI used [\"ImageId\", \"LabelId\",\"PixelCount\", \"Confidence\", \"EncodedPixels\"] as csv’s header and the real data use this order([“ImageId\", \"LabelId\", \"Confidence\", \"PixelCount\", \"EncodedPixels”]), I could got a very high score(0.31),but when I use [“ImageId\", \"LabelId\", \"Confidence\", \"PixelCount\", \"EncodedPixels”] as header and real data’s order, I got a worse score(0.11).</p>",
      "rawMarkdown": "Thanks LiBin for Helping us fixed a new bug!\n\n&nbsp; &nbsp; &nbsp; I fixed a new bug. If user's&nbsp;PixelCount in&nbsp;submission is wrong, the result will be hacked.\n&nbsp; &nbsp; &nbsp;&nbsp;it because&nbsp;overlap = pred.intersection / (gt.pixelCount +&nbsp;pred.pixelCount&nbsp;- pred.intersection);\n&nbsp; &nbsp; &nbsp; It means if submission set&nbsp;all pixelCount = 0.&nbsp;&nbsp;the overlap will be much higher than it should be.&nbsp;\n\n&nbsp; &nbsp; &nbsp; I have changed code to let program count pixelCount by itself, instead of using user's.&nbsp;\n&nbsp; &nbsp; &nbsp; The function is&nbsp;&nbsp;DecodePridictionCVS_Kaggle(CVPRAutoDrivingAveragePrecision.Submission _submission)\n\nLiBin’s report:\nI used [\"ImageId\", \"LabelId\",\"PixelCount\", \"Confidence\", \"EncodedPixels\"] as csv’s header and the real data use this order([“ImageId\", \"LabelId\", \"Confidence\", \"PixelCount\", \"EncodedPixels”]), I could got a very high score(0.31),but when I use [“ImageId\", \"LabelId\", \"Confidence\", \"PixelCount\", \"EncodedPixels”] as header and real data’s order, I got a worse score(0.11).",
      "votes": null
    },
    {
      "id": "340160",
      "postDate": "06/08/2018 13:50:29",
      "content": "<p>but why confidence set to 1, the result will be high ? </p>",
      "rawMarkdown": "but why confidence set to 1, the result will be high ?",
      "votes": null
    },
    {
      "id": "340162",
      "postDate": "06/08/2018 13:53:38",
      "content": "<p>We sincerely admire Li Bin’s integrity! Because Li Bin’s score was Top1 at the time, his score was greater than 0.3. However, he still reported this bug so that his score would be reduced to less than 0.15. I will thank you again and look forward to seeing us at CVPR.</p>",
      "rawMarkdown": "We sincerely admire Li Bin’s integrity! Because Li Bin’s score was Top1 at the time, his score was greater than 0.3. However, he still reported this bug so that his score would be reduced to less than 0.15. I will thank you again and look forward to seeing us at CVPR.",
      "votes": null
    },
    {
      "id": "340164",
      "postDate": "06/08/2018 13:54:04",
      "content": "<p>I will continue to check this :P</p>",
      "rawMarkdown": "I will continue to check this :P",
      "votes": null
    },
    {
      "id": "340214",
      "postDate": "06/08/2018 16:11:28",
      "content": "<p>Thanks to Li Bin and admire him!!\nTwo questions: 1. Does that mean all past scores need to be recalculated? 2. I assume recounting pixels take some server time, and does this mean some previous successfully submitted results may not passing the time limit now?</p>",
      "rawMarkdown": "Thanks to Li Bin and admire him!!\nTwo questions: 1. Does that mean all past scores need to be recalculated? 2. I assume recounting pixels take some server time, and does this mean some previous successfully submitted results may not passing the time limit now?",
      "votes": null
    },
    {
      "id": "340235",
      "postDate": "06/08/2018 17:16:22",
      "content": "<p>代码放出来前不应该充分的测试吗？感觉有些低级错误不应该犯啊。很无语。希望贵公司好好努力吧。</p>",
      "rawMarkdown": "代码放出来前不应该充分的测试吗？感觉有些低级错误不应该犯啊。很无语。希望贵公司好好努力吧。",
      "votes": null
    },
    {
      "id": "340257",
      "postDate": "06/08/2018 18:29:21",
      "content": "<p>Same questions here. I did the test for me: pulling down my best submission and re-submitting it now times out...</p>",
      "rawMarkdown": "Same questions here. I did the test for me: pulling down my best submission and re-submitting it now times out...",
      "votes": null
    },
    {
      "id": "340305",
      "postDate": "06/08/2018 21:01:17",
      "content": "<p>LMAO, 3 days to go :D\nWhen is the leaderboard​ going to be fixed with respect to all the bugs found? </p>",
      "rawMarkdown": "LMAO, 3 days to go :D\nWhen is the leaderboard​ going to be fixed with respect to all the bugs found?",
      "votes": null
    },
    {
      "id": "340609",
      "postDate": "06/09/2018 18:59:03",
      "content": "<p>We'll be kicking off a rescore soon to re-evaluate old submissions. </p>",
      "rawMarkdown": "We'll be kicking off a rescore soon to re-evaluate old submissions.",
      "votes": null
    },
    {
      "id": "340613",
      "postDate": "06/09/2018 19:16:38",
      "content": "<p>I guess this is happening right now? The leaderboard looks messed up (plus my submission is taking a long time without timing out)</p>",
      "rawMarkdown": "I guess this is happening right now? The leaderboard looks messed up (plus my submission is taking a long time without timing out)",
      "votes": null
    },
    {
      "id": "340645",
      "postDate": "06/09/2018 22:10:42",
      "content": "<p>Will we get additional submissions? As I still have two submissions left \"today\" and I currently can't use them</p>",
      "rawMarkdown": "Will we get additional submissions? As I still have two submissions left \"today\" and I currently can't use them",
      "votes": null
    },
    {
      "id": "340681",
      "postDate": "06/10/2018 02:16:13",
      "content": "<p>Yes, it's still ongoing. Your current submissions won't get affected since the new submissions have higher priority than the rescoring submissions. </p>",
      "rawMarkdown": "Yes, it's still ongoing. Your current submissions won't get affected since the new submissions have higher priority than the rescoring submissions.",
      "votes": null
    },
    {
      "id": "340682",
      "postDate": "06/10/2018 02:26:04",
      "content": "<p>Not Right， I submit new submission but cannot be calculated。</p>",
      "rawMarkdown": "Not Right， I submit new submission but cannot be calculated。",
      "votes": null
    },
    {
      "id": "340683",
      "postDate": "06/10/2018 02:28:56",
      "content": "<p>Yes, the priorities are messed up, my new submission has been stuck for 8h now. This is why I was asking for additional submissions. I could and still can not use my daily quota of submissions (lost two yesterday)</p>",
      "rawMarkdown": "Yes, the priorities are messed up, my new submission has been stuck for 8h now. This is why I was asking for additional submissions. I could and still can not use my daily quota of submissions (lost two yesterday)",
      "votes": null
    },
    {
      "id": "340693",
      "postDate": "06/10/2018 03:30:51",
      "content": "<p>Sorry for the inconvenience @Michael Heinzer and @Tommy.Zhuang, these last minute bug fixes are putting quite heavy stress on the server. I'll manually error out your submissions that are stuck so you can make new submissions. I'll also increase the daily submission limit to 5/day. </p>",
      "rawMarkdown": "Sorry for the inconvenience @Michael Heinzer and @Tommy.Zhuang, these last minute bug fixes are putting quite heavy stress on the server. I'll manually error out your submissions that are stuck so you can make new submissions. I'll also increase the daily submission limit to 5/day.",
      "votes": null
    },
    {
      "id": "340694",
      "postDate": "06/10/2018 03:36:28",
      "content": "<p>It means all submissions that I submitted are invalid ?! </p>",
      "rawMarkdown": "It means all submissions that I submitted are invalid ?!",
      "votes": null
    },
    {
      "id": "340700",
      "postDate": "06/10/2018 03:50:52",
      "content": "<p>Perfect, thanks! </p>",
      "rawMarkdown": "Perfect, thanks!",
      "votes": null
    },
    {
      "id": "340715",
      "postDate": "06/10/2018 05:31:53",
      "content": "<p>@Tommy.Zhuang, no, it means they will be re-processed. Please let me know if you are still stuck. </p>",
      "rawMarkdown": "Tommy.Zhuang, no, it means they will be re-processed. Please let me know if you are still stuck.",
      "votes": null
    },
    {
      "id": "340717",
      "postDate": "06/10/2018 05:37:49",
      "content": "<p>I am still stucked !!!!</p>",
      "rawMarkdown": "I am still stucked !!!!",
      "votes": null
    },
    {
      "id": "340725",
      "postDate": "06/10/2018 06:07:34",
      "content": "<p>I submit result, but it cannot be shown !!!! </p>",
      "rawMarkdown": "I submit result, but it cannot be shown !!!!",
      "votes": null
    },
    {
      "id": "340728",
      "postDate": "06/10/2018 06:16:14",
      "content": "<p>Still stuck too</p>",
      "rawMarkdown": "Still stuck too",
      "votes": null
    },
    {
      "id": "340876",
      "postDate": "06/10/2018 14:45:08",
      "content": "<p>Pls  give me a special contribution award~</p>",
      "rawMarkdown": "Pls  give me a special contribution award~",
      "votes": null
    },
    {
      "id": "340881",
      "postDate": "06/10/2018 14:50:13",
      "content": "<p>bug report award !!!</p>",
      "rawMarkdown": "bug report award !!!",
      "votes": null
    },
    {
      "id": "340900",
      "postDate": "06/10/2018 15:28:20",
      "content": "<p>I wonder after this correction and re-scoring, did anyone submit their result successfully when the size of the file is larger than 35MB? </p>",
      "rawMarkdown": "I wonder after this correction and re-scoring, did anyone submit their result successfully when the size of the file is larger than 35MB?",
      "votes": null
    },
    {
      "id": "340903",
      "postDate": "06/10/2018 15:41:38",
      "content": "<p>I think, none</p>",
      "rawMarkdown": "I think, none",
      "votes": null
    },
    {
      "id": "340907",
      "postDate": "06/10/2018 15:49:30",
      "content": "<p>I haven't been able to successfully submit any file yet.</p>",
      "rawMarkdown": "I haven't been able to successfully submit any file yet.",
      "votes": null
    },
    {
      "id": "341549",
      "postDate": "06/11/2018 19:20:03",
      "content": "<p>It seems resolved for me.</p>",
      "rawMarkdown": "It seems resolved for me.",
      "votes": null
    },
    {
      "id": "341586",
      "postDate": "06/11/2018 20:41:08",
      "content": "<p>Yeah, it was resolved once the leaderboard was recalculated</p>",
      "rawMarkdown": "Yeah, it was resolved once the leaderboard was recalculated",
      "votes": null
    },
    {
      "id": "341623",
      "postDate": "06/11/2018 23:52:18",
      "content": "<p>Finally, it works on the last day. Good luck to everyone. </p>",
      "rawMarkdown": "Finally, it works on the last day. Good luck to everyone.",
      "votes": null
    },
    {
      "id": "343807",
      "postDate": "06/16/2018 07:06:19",
      "content": "<p>Dear sx,</p>\n\n<p>I am writing this message to congratulating you and your team on the second place in the CVPR 2018 WAD Video Segmentation Challenge.</p>\n\n<p>As you may know, our workshop will be held on June 18th and June 22nd 2018 at Salt Lake Place Center in Utah. We have invited 14 leading scholars to give speeches at the workshop. We also would like to invite you to attend the workshop, especially the Awards and Winner Presentation between 9:30AM and 10:20AM on June 22nd. We have prepared nice looking certificates for all the winners. :) </p>\n\n<p>Moreover, we are selecting four winning teams to present their works (around 10 minutes each). Please let us know if you could attend the workshop (especially on June 22nd) so that we could finalize the workshop procedure. </p>\n\n<p>Congratulation again and wish you all the best with success in academic and professional pursuits.</p>\n\n<p>Sincerely,</p>\n\n<p>Xinyu Huang</p>\n\n<p>Baidu Research</p>",
      "rawMarkdown": "Dear sx,\n\nI am writing this message to congratulating you and your team on the second place in the CVPR 2018 WAD Video Segmentation Challenge.\n\nAs you may know, our workshop will be held on June 18th and June 22nd 2018 at Salt Lake Place Center in Utah. We have invited 14 leading scholars to give speeches at the workshop. We also would like to invite you to attend the workshop, especially the Awards and Winner Presentation between 9:30AM and 10:20AM on June 22nd. We have prepared nice looking certificates for all the winners. :) \n\nMoreover, we are selecting four winning teams to present their works (around 10 minutes each). Please let us know if you could attend the workshop (especially on June 22nd) so that we could finalize the workshop procedure. \n\nCongratulation again and wish you all the best with success in academic and professional pursuits.\n\nSincerely,\n\nXinyu Huang\n\nBaidu Research",
      "votes": null
    },
    {
      "id": "344926",
      "postDate": "06/18/2018 23:53:11",
      "content": "<p>Hi Xinyu,\nThanks. Unfortunately, I am on holiday and not able to present my work in the workshop. However, I am preparing the report and will share with the community.\nBest</p>",
      "rawMarkdown": "Hi Xinyu,\nThanks. Unfortunately, I am on holiday and not able to present my work in the workshop. However, I am preparing the report and will share with the community.\nBest",
      "votes": null
    },
    {
      "id": "345412",
      "postDate": "06/19/2018 20:45:01",
      "content": "<p>sx, could you please provide your contact information, company or personal information to us? we will announce winners on June 22nd. my email is huangxinyu01@baidu.com, thanks!</p>",
      "rawMarkdown": "sx, could you please provide your contact information, company or personal information to us? we will announce winners on June 22nd. my email is huangxinyu01@baidu.com, thanks!",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 340160,
      "author_name": "zhuangyq",
      "author_url": "",
      "post_date": "06/08/2018 13:50:29",
      "content": "<p>but why confidence set to 1, the result will be high ? </p>",
      "votes": null,
      "replies": [
        {
          "id": 340164,
          "author_name": "shaochenye",
          "author_url": "",
          "post_date": "06/08/2018 13:54:04",
          "content": "<p>I will continue to check this :P</p>",
          "votes": null,
          "replies": []
        }
      ]
    },
    {
      "id": 340162,
      "author_name": "shaochenye",
      "author_url": "",
      "post_date": "06/08/2018 13:53:38",
      "content": "<p>We sincerely admire Li Bin’s integrity! Because Li Bin’s score was Top1 at the time, his score was greater than 0.3. However, he still reported this bug so that his score would be reduced to less than 0.15. I will thank you again and look forward to seeing us at CVPR.</p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 340214,
      "author_name": "woodywang",
      "author_url": "",
      "post_date": "06/08/2018 16:11:28",
      "content": "<p>Thanks to Li Bin and admire him!!\nTwo questions: 1. Does that mean all past scores need to be recalculated? 2. I assume recounting pixels take some server time, and does this mean some previous successfully submitted results may not passing the time limit now?</p>",
      "votes": null,
      "replies": [
        {
          "id": 340257,
          "author_name": "nvadlr",
          "author_url": "",
          "post_date": "06/08/2018 18:29:21",
          "content": "<p>Same questions here. I did the test for me: pulling down my best submission and re-submitting it now times out...</p>",
          "votes": null,
          "replies": []
        }
      ]
    },
    {
      "id": 340235,
      "author_name": "andeyeluguo",
      "author_url": "",
      "post_date": "06/08/2018 17:16:22",
      "content": "<p>代码放出来前不应该充分的测试吗？感觉有些低级错误不应该犯啊。很无语。希望贵公司好好努力吧。</p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 340305,
      "author_name": "azakhtyamov",
      "author_url": "",
      "post_date": "06/08/2018 21:01:17",
      "content": "<p>LMAO, 3 days to go :D\nWhen is the leaderboard​ going to be fixed with respect to all the bugs found? </p>",
      "votes": null,
      "replies": [
        {
          "id": 340609,
          "author_name": "wendykan",
          "author_url": "",
          "post_date": "06/09/2018 18:59:03",
          "content": "<p>We'll be kicking off a rescore soon to re-evaluate old submissions. </p>",
          "votes": null,
          "replies": []
        },
        {
          "id": 340613,
          "author_name": "michaelheinzer",
          "author_url": "",
          "post_date": "06/09/2018 19:16:38",
          "content": "<p>I guess this is happening right now? The leaderboard looks messed up (plus my submission is taking a long time without timing out)</p>",
          "votes": null,
          "replies": []
        },
        {
          "id": 340645,
          "author_name": "michaelheinzer",
          "author_url": "",
          "post_date": "06/09/2018 22:10:42",
          "content": "<p>Will we get additional submissions? As I still have two submissions left \"today\" and I currently can't use them</p>",
          "votes": null,
          "replies": []
        },
        {
          "id": 340681,
          "author_name": "wendykan",
          "author_url": "",
          "post_date": "06/10/2018 02:16:13",
          "content": "<p>Yes, it's still ongoing. Your current submissions won't get affected since the new submissions have higher priority than the rescoring submissions. </p>",
          "votes": null,
          "replies": []
        },
        {
          "id": 340682,
          "author_name": "zhuangyq",
          "author_url": "",
          "post_date": "06/10/2018 02:26:04",
          "content": "<p>Not Right， I submit new submission but cannot be calculated。</p>",
          "votes": null,
          "replies": []
        },
        {
          "id": 340683,
          "author_name": "michaelheinzer",
          "author_url": "",
          "post_date": "06/10/2018 02:28:56",
          "content": "<p>Yes, the priorities are messed up, my new submission has been stuck for 8h now. This is why I was asking for additional submissions. I could and still can not use my daily quota of submissions (lost two yesterday)</p>",
          "votes": null,
          "replies": []
        },
        {
          "id": 340693,
          "author_name": "wendykan",
          "author_url": "",
          "post_date": "06/10/2018 03:30:51",
          "content": "<p>Sorry for the inconvenience @Michael Heinzer and @Tommy.Zhuang, these last minute bug fixes are putting quite heavy stress on the server. I'll manually error out your submissions that are stuck so you can make new submissions. I'll also increase the daily submission limit to 5/day. </p>",
          "votes": null,
          "replies": []
        },
        {
          "id": 340694,
          "author_name": "zhuangyq",
          "author_url": "",
          "post_date": "06/10/2018 03:36:28",
          "content": "<p>It means all submissions that I submitted are invalid ?! </p>",
          "votes": null,
          "replies": []
        },
        {
          "id": 340700,
          "author_name": "michaelheinzer",
          "author_url": "",
          "post_date": "06/10/2018 03:50:52",
          "content": "<p>Perfect, thanks! </p>",
          "votes": null,
          "replies": []
        },
        {
          "id": 340715,
          "author_name": "wendykan",
          "author_url": "",
          "post_date": "06/10/2018 05:31:53",
          "content": "<p>@Tommy.Zhuang, no, it means they will be re-processed. Please let me know if you are still stuck. </p>",
          "votes": null,
          "replies": []
        },
        {
          "id": 340717,
          "author_name": "zhuangyq",
          "author_url": "",
          "post_date": "06/10/2018 05:37:49",
          "content": "<p>I am still stucked !!!!</p>",
          "votes": null,
          "replies": []
        },
        {
          "id": 340725,
          "author_name": "zhuangyq",
          "author_url": "",
          "post_date": "06/10/2018 06:07:34",
          "content": "<p>I submit result, but it cannot be shown !!!! </p>",
          "votes": null,
          "replies": []
        },
        {
          "id": 340728,
          "author_name": "michaelheinzer",
          "author_url": "",
          "post_date": "06/10/2018 06:16:14",
          "content": "<p>Still stuck too</p>",
          "votes": null,
          "replies": []
        },
        {
          "id": 341549,
          "author_name": "amitkumarjaiswal",
          "author_url": "",
          "post_date": "06/11/2018 19:20:03",
          "content": "<p>It seems resolved for me.</p>",
          "votes": null,
          "replies": []
        },
        {
          "id": 341586,
          "author_name": "michaelheinzer",
          "author_url": "",
          "post_date": "06/11/2018 20:41:08",
          "content": "<p>Yeah, it was resolved once the leaderboard was recalculated</p>",
          "votes": null,
          "replies": []
        }
      ]
    },
    {
      "id": 340876,
      "author_name": "the16throute",
      "author_url": "",
      "post_date": "06/10/2018 14:45:08",
      "content": "<p>Pls  give me a special contribution award~</p>",
      "votes": null,
      "replies": [
        {
          "id": 340881,
          "author_name": "zhuangyq",
          "author_url": "",
          "post_date": "06/10/2018 14:50:13",
          "content": "<p>bug report award !!!</p>",
          "votes": null,
          "replies": []
        }
      ]
    },
    {
      "id": 340900,
      "author_name": "shanxu",
      "author_url": "",
      "post_date": "06/10/2018 15:28:20",
      "content": "<p>I wonder after this correction and re-scoring, did anyone submit their result successfully when the size of the file is larger than 35MB? </p>",
      "votes": null,
      "replies": [
        {
          "id": 340903,
          "author_name": "tianmaliuxing",
          "author_url": "",
          "post_date": "06/10/2018 15:41:38",
          "content": "<p>I think, none</p>",
          "votes": null,
          "replies": []
        },
        {
          "id": 340907,
          "author_name": "woodywang",
          "author_url": "",
          "post_date": "06/10/2018 15:49:30",
          "content": "<p>I haven't been able to successfully submit any file yet.</p>",
          "votes": null,
          "replies": []
        },
        {
          "id": 341623,
          "author_name": "shanxu",
          "author_url": "",
          "post_date": "06/11/2018 23:52:18",
          "content": "<p>Finally, it works on the last day. Good luck to everyone. </p>",
          "votes": null,
          "replies": []
        },
        {
          "id": 343807,
          "author_name": "huangxinyu01",
          "author_url": "",
          "post_date": "06/16/2018 07:06:19",
          "content": "<p>Dear sx,</p>\n\n<p>I am writing this message to congratulating you and your team on the second place in the CVPR 2018 WAD Video Segmentation Challenge.</p>\n\n<p>As you may know, our workshop will be held on June 18th and June 22nd 2018 at Salt Lake Place Center in Utah. We have invited 14 leading scholars to give speeches at the workshop. We also would like to invite you to attend the workshop, especially the Awards and Winner Presentation between 9:30AM and 10:20AM on June 22nd. We have prepared nice looking certificates for all the winners. :) </p>\n\n<p>Moreover, we are selecting four winning teams to present their works (around 10 minutes each). Please let us know if you could attend the workshop (especially on June 22nd) so that we could finalize the workshop procedure. </p>\n\n<p>Congratulation again and wish you all the best with success in academic and professional pursuits.</p>\n\n<p>Sincerely,</p>\n\n<p>Xinyu Huang</p>\n\n<p>Baidu Research</p>",
          "votes": null,
          "replies": []
        },
        {
          "id": 344926,
          "author_name": "shanxu",
          "author_url": "",
          "post_date": "06/18/2018 23:53:11",
          "content": "<p>Hi Xinyu,\nThanks. Unfortunately, I am on holiday and not able to present my work in the workshop. However, I am preparing the report and will share with the community.\nBest</p>",
          "votes": null,
          "replies": []
        },
        {
          "id": 345412,
          "author_name": "huangxinyu01",
          "author_url": "",
          "post_date": "06/19/2018 20:45:01",
          "content": "<p>sx, could you please provide your contact information, company or personal information to us? we will announce winners on June 22nd. my email is huangxinyu01@baidu.com, thanks!</p>",
          "votes": null,
          "replies": []
        }
      ]
    }
  ],
  "raw_markdown_by_id": {
    "340159": "Thanks LiBin for Helping us fixed a new bug!\n\n&nbsp; &nbsp; &nbsp; I fixed a new bug. If user's&nbsp;PixelCount in&nbsp;submission is wrong, the result will be hacked.\n&nbsp; &nbsp; &nbsp;&nbsp;it because&nbsp;overlap = pred.intersection / (gt.pixelCount +&nbsp;pred.pixelCount&nbsp;- pred.intersection);\n&nbsp; &nbsp; &nbsp; It means if submission set&nbsp;all pixelCount = 0.&nbsp;&nbsp;the overlap will be much higher than it should be.&nbsp;\n\n&nbsp; &nbsp; &nbsp; I have changed code to let program count pixelCount by itself, instead of using user's.&nbsp;\n&nbsp; &nbsp; &nbsp; The function is&nbsp;&nbsp;DecodePridictionCVS_Kaggle(CVPRAutoDrivingAveragePrecision.Submission _submission)\n\nLiBin’s report:\nI used [\"ImageId\", \"LabelId\",\"PixelCount\", \"Confidence\", \"EncodedPixels\"] as csv’s header and the real data use this order([“ImageId\", \"LabelId\", \"Confidence\", \"PixelCount\", \"EncodedPixels”]), I could got a very high score(0.31),but when I use [“ImageId\", \"LabelId\", \"Confidence\", \"PixelCount\", \"EncodedPixels”] as header and real data’s order, I got a worse score(0.11).",
    "340160": "but why confidence set to 1, the result will be high ?",
    "340162": "We sincerely admire Li Bin’s integrity! Because Li Bin’s score was Top1 at the time, his score was greater than 0.3. However, he still reported this bug so that his score would be reduced to less than 0.15. I will thank you again and look forward to seeing us at CVPR.",
    "340164": "I will continue to check this :P",
    "340214": "Thanks to Li Bin and admire him!!\nTwo questions: 1. Does that mean all past scores need to be recalculated? 2. I assume recounting pixels take some server time, and does this mean some previous successfully submitted results may not passing the time limit now?",
    "340235": "代码放出来前不应该充分的测试吗？感觉有些低级错误不应该犯啊。很无语。希望贵公司好好努力吧。",
    "340257": "Same questions here. I did the test for me: pulling down my best submission and re-submitting it now times out...",
    "340305": "LMAO, 3 days to go :D\nWhen is the leaderboard​ going to be fixed with respect to all the bugs found?",
    "340609": "We'll be kicking off a rescore soon to re-evaluate old submissions.",
    "340613": "I guess this is happening right now? The leaderboard looks messed up (plus my submission is taking a long time without timing out)",
    "340645": "Will we get additional submissions? As I still have two submissions left \"today\" and I currently can't use them",
    "340681": "Yes, it's still ongoing. Your current submissions won't get affected since the new submissions have higher priority than the rescoring submissions.",
    "340682": "Not Right， I submit new submission but cannot be calculated。",
    "340683": "Yes, the priorities are messed up, my new submission has been stuck for 8h now. This is why I was asking for additional submissions. I could and still can not use my daily quota of submissions (lost two yesterday)",
    "340693": "Sorry for the inconvenience @Michael Heinzer and @Tommy.Zhuang, these last minute bug fixes are putting quite heavy stress on the server. I'll manually error out your submissions that are stuck so you can make new submissions. I'll also increase the daily submission limit to 5/day.",
    "340694": "It means all submissions that I submitted are invalid ?!",
    "340700": "Perfect, thanks!",
    "340715": "Tommy.Zhuang, no, it means they will be re-processed. Please let me know if you are still stuck.",
    "340717": "I am still stucked !!!!",
    "340725": "I submit result, but it cannot be shown !!!!",
    "340728": "Still stuck too",
    "340876": "Pls  give me a special contribution award~",
    "340881": "bug report award !!!",
    "340900": "I wonder after this correction and re-scoring, did anyone submit their result successfully when the size of the file is larger than 35MB?",
    "340903": "I think, none",
    "340907": "I haven't been able to successfully submit any file yet.",
    "341549": "It seems resolved for me.",
    "341586": "Yeah, it was resolved once the leaderboard was recalculated",
    "341623": "Finally, it works on the last day. Good luck to everyone.",
    "343807": "Dear sx,\n\nI am writing this message to congratulating you and your team on the second place in the CVPR 2018 WAD Video Segmentation Challenge.\n\nAs you may know, our workshop will be held on June 18th and June 22nd 2018 at Salt Lake Place Center in Utah. We have invited 14 leading scholars to give speeches at the workshop. We also would like to invite you to attend the workshop, especially the Awards and Winner Presentation between 9:30AM and 10:20AM on June 22nd. We have prepared nice looking certificates for all the winners. :) \n\nMoreover, we are selecting four winning teams to present their works (around 10 minutes each). Please let us know if you could attend the workshop (especially on June 22nd) so that we could finalize the workshop procedure. \n\nCongratulation again and wish you all the best with success in academic and professional pursuits.\n\nSincerely,\n\nXinyu Huang\n\nBaidu Research",
    "344926": "Hi Xinyu,\nThanks. Unfortunately, I am on holiday and not able to present my work in the workshop. However, I am preparing the report and will share with the community.\nBest",
    "345412": "sx, could you please provide your contact information, company or personal information to us? we will announce winners on June 22nd. my email is huangxinyu01@baidu.com, thanks!"
  },
  "source": "meta"
}