{
  "id": 138639,
  "title": "REMINDER: Select your Final 2 Submissions by 3/31",
  "url": "/competitions/deepfake-detection-challenge/discussion/138639",
  "author_name": "Julia Elliott",
  "post_date": "2020-03-25T19:07:16.485000",
  "votes": 14,
  "comment_count": 16,
  "views": 0,
  "content": "<p>This is a reminder that the deadline for submission AND to select the 2 submissions that will be re-run by the host on the private test set is <strong>March 31st 2020 at 11:59pm UTC</strong>. To be clear, we are <strong>only re-running selected submissions</strong>. You will <strong>not</strong> receive a final private score on your other submissions. </p>\n\n<p>A few things to keep in mind (none of this should be new information):</p>\n\n<ol>\n<li>Ensure your code is responsive to a completely new test set and id's. Hard-coding of values will fail. The way this competition was designed -- with a hidden public leaderboard test set -- should have encouraged this to be the case, if you were able to submit and receive a public score without error.</li>\n<li>As stated in the <a href=\"https://w.kaggle.com/c/deepfake-detection-challenge/overview/getting-started\">getting started</a> page, the time constraints applied to the private test run will scale proportionally to the size of that test set.</li>\n<li>The file structure and naming of the files (<code>test_videos</code>, <code>sample_submission.csv</code>) currently available on the Data page and file paths expected within your notebook sessions will remain the same. The difference will be in the <em>contents</em> of those files (i.e. the videos themselves in test_videos will be swapped out for the actual private test videos).</li>\n<li>If you do not make a submission selections, the auto-selection will default to your highest public-leaderboard-scoring submission's notebook(s). This may or may not represent what you expect. So it's always safest to make the manual selections.</li>\n<li>If you are part of a team, only the team leader may make the final selections.</li>\n<li>Don't risk deleting your selected notebooks <em>or</em> associated external datasets before the final private leaderboard results are confirmed. If we're unable to access those notebooks/datasets during re-runs, you will not be able to receive a score on that notebook.</li>\n</ol>\n\n<p>Best of luck in the final days of this competition!</p>",
  "messages": [
    {
      "id": 786293,
      "postDate": "2020-03-25T19:07:16.487Z",
      "content": "<p>This is a reminder that the deadline for submission AND to select the 2 submissions that will be re-run by the host on the private test set is <strong>March 31st 2020 at 11:59pm UTC</strong>. To be clear, we are <strong>only re-running selected submissions</strong>. You will <strong>not</strong> receive a final private score on your other submissions. </p>\n\n<p>A few things to keep in mind (none of this should be new information):</p>\n\n<ol>\n<li>Ensure your code is responsive to a completely new test set and id's. Hard-coding of values will fail. The way this competition was designed -- with a hidden public leaderboard test set -- should have encouraged this to be the case, if you were able to submit and receive a public score without error.</li>\n<li>As stated in the <a href=\"https://w.kaggle.com/c/deepfake-detection-challenge/overview/getting-started\">getting started</a> page, the time constraints applied to the private test run will scale proportionally to the size of that test set.</li>\n<li>The file structure and naming of the files (<code>test_videos</code>, <code>sample_submission.csv</code>) currently available on the Data page and file paths expected within your notebook sessions will remain the same. The difference will be in the <em>contents</em> of those files (i.e. the videos themselves in test_videos will be swapped out for the actual private test videos).</li>\n<li>If you do not make a submission selections, the auto-selection will default to your highest public-leaderboard-scoring submission's notebook(s). This may or may not represent what you expect. So it's always safest to make the manual selections.</li>\n<li>If you are part of a team, only the team leader may make the final selections.</li>\n<li>Don't risk deleting your selected notebooks <em>or</em> associated external datasets before the final private leaderboard results are confirmed. If we're unable to access those notebooks/datasets during re-runs, you will not be able to receive a score on that notebook.</li>\n</ol>\n\n<p>Best of luck in the final days of this competition!</p>",
      "rawMarkdown": "This is a reminder that the deadline for submission AND to select the 2 submissions that will be re-run by the host on the private test set is **March 31st 2020 at 11:59pm UTC**. To be clear, we are **only re-running selected submissions**. You will **not** receive a final private score on your other submissions. \n\nA few things to keep in mind (none of this should be new information):\n\n1. Ensure your code is responsive to a completely new test set and id's. Hard-coding of values will fail. The way this competition was designed -- with a hidden public leaderboard test set -- should have encouraged this to be the case, if you were able to submit and receive a public score without error.\n2. As stated in the [getting started](https://w.kaggle.com/c/deepfake-detection-challenge/overview/getting-started) page, the time constraints applied to the private test run will scale proportionally to the size of that test set.\n3. The file structure and naming of the files (`test_videos`, `sample_submission.csv`) currently available on the Data page and file paths expected within your notebook sessions will remain the same. The difference will be in the *contents* of those files (i.e. the videos themselves in test_videos will be swapped out for the actual private test videos).\n4. If you do not make a submission selections, the auto-selection will default to your highest public-leaderboard-scoring submission's notebook(s). This may or may not represent what you expect. So it's always safest to make the manual selections.\n5. If you are part of a team, only the team leader may make the final selections.\n6. Don't risk deleting your selected notebooks *or* associated external datasets before the final private leaderboard results are confirmed. If we're unable to access those notebooks/datasets during re-runs, you will not be able to receive a score on that notebook.\n\nBest of luck in the final days of this competition!",
      "votes": 14
    },
    {
      "id": 788269,
      "postDate": "2020-03-27T14:28:02.733Z",
      "content": "<p>Hi Julia,\n I was also affected by the covid and could not work for about 4-5 weeks being in a \"hotspot\" location. It would be great to get a few weeks of extension if this is agreeable with the others. Thanks <a href=\"/juliaelliott\">@juliaelliott</a> </p>\n\n<p>Regards.</p>",
      "rawMarkdown": "Hi Julia,\n I was also affected by the covid and could not work for about 4-5 weeks being in a \"hotspot\" location. It would be great to get a few weeks of extension if this is agreeable with the others. Thanks @juliaelliott \n\nRegards.",
      "votes": 4
    },
    {
      "id": 787754,
      "postDate": "2020-03-27T03:47:56.030Z",
      "content": "<p>I would like to ask for a 2 week extension. We have a pandemic that most likely affected the performance of many of us. We need additional time. If possible, can the dead line be extended. It's quite tight for many of us as it is.</p>",
      "rawMarkdown": "I would like to ask for a 2 week extension. We have a pandemic that most likely affected the performance of many of us. We need additional time. If possible, can the dead line be extended. It's quite tight for many of us as it is.\n",
      "votes": 1
    },
    {
      "id": 792565,
      "postDate": "2020-03-31T10:34:48.077Z",
      "rawMarkdown": "",
      "votes": 1,
      "replies": [
        {
          "id": 792604,
          "postDate": "2020-03-31T11:13:30.773Z",
          "content": "<p><a href=\"/adamrr\">@adamrr</a>  A lot of us (115+) had been asking for Extension on the other thread (<a href=\"https://www.kaggle.com/c/deepfake-detection-challenge/discussion/135631\">https://www.kaggle.com/c/deepfake-detection-challenge/discussion/135631</a>), but all our requests were denied 😪 </p>",
          "rawMarkdown": "@adamrr  A lot of us (115+) had been asking for Extension on the other thread (https://www.kaggle.com/c/deepfake-detection-challenge/discussion/135631), but all our requests were denied 😪 ",
          "votes": 1
        }
      ]
    },
    {
      "id": 791054,
      "postDate": "2020-03-30T04:33:11.640Z",
      "content": "<p>Hi Julia - I've recently tried to make submissions but there were errors during the process - at the moment those submissions are counting towards my total allowed submissions per day and I am not able to try submitting again.  Based on my validation score I really think I have a chance of being competitive - I only have 4 submissions left given the timing towards the deadline and I'm quite nervous that those 4 submissions chances will result in errors as well and prevent me from making any submission at all.  Are there any workarounds regarding submission errors preventing me from trying to resubmit?  At the moment it's very difficult to troubleshoot the issue without \"risking it all.\"  Thanks in advance for your help.  </p>",
      "rawMarkdown": "Hi Julia - I've recently tried to make submissions but there were errors during the process - at the moment those submissions are counting towards my total allowed submissions per day and I am not able to try submitting again.  Based on my validation score I really think I have a chance of being competitive - I only have 4 submissions left given the timing towards the deadline and I'm quite nervous that those 4 submissions chances will result in errors as well and prevent me from making any submission at all.  Are there any workarounds regarding submission errors preventing me from trying to resubmit?  At the moment it's very difficult to troubleshoot the issue without \"risking it all.\"  Thanks in advance for your help.  ",
      "votes": 1
    },
    {
      "id": 793271,
      "postDate": "2020-03-31T23:18:53.577Z",
      "content": "<p>Could you please give us several hours for selecting two submissions after UTC 23.59???? I know we must submit our last submission before 11.59 pm UTC. But the two submissions are still running, we are not able to see whether they are successfully submitted and see the scores. Also, we cannot select these two!</p>",
      "rawMarkdown": "Could you please give us several hours for selecting two submissions after UTC 23.59???? I know we must submit our last submission before 11.59 pm UTC. But the two submissions are still running, we are not able to see whether they are successfully submitted and see the scores. Also, we cannot select these two!",
      "replies": [
        {
          "id": 793289,
          "postDate": "2020-03-31T23:32:04.730Z",
          "content": "<p>Please see this thread: <a href=\"https://www.kaggle.com/c/deepfake-detection-challenge/discussion/139986\">https://www.kaggle.com/c/deepfake-detection-challenge/discussion/139986</a></p>\n\n<p>It is a pity that you have such situation, but rules are equal for everyone. So if you wanted to make submit in the last day, you should have left maximum 9 hours for it till the deadline. If you didn't, then it's your fault and lesson for the next code competitions, sorry)</p>",
          "rawMarkdown": "Please see this thread: https://www.kaggle.com/c/deepfake-detection-challenge/discussion/139986\n\nIt is a pity that you have such situation, but rules are equal for everyone. So if you wanted to make submit in the last day, you should have left maximum 9 hours for it till the deadline. If you didn't, then it's your fault and lesson for the next code competitions, sorry)"
        }
      ]
    },
    {
      "id": 793109,
      "postDate": "2020-03-31T20:00:28.177Z",
      "content": "<p>There seems to be a lot of submission error. The Notebook system is out of sync with data.  Some submissions look impossible!.</p>\n\n<p>Really not a good time to do those changes in the final day of this competition. </p>\n\n<p>I think many will be affected by this.</p>",
      "rawMarkdown": "There seems to be a lot of submission error. The Notebook system is out of sync with data.  Some submissions look impossible!.\n\nReally not a good time to do those changes in the final day of this competition. \n\nI think many will be affected by this."
    },
    {
      "id": 793015,
      "postDate": "2020-03-31T18:14:25.867Z",
      "content": "<p>HI Julia:\n   Not a good time to change the UI on us this morning!.  We had a commit button , now it's save a new version, I wasn't able to tell what this is! .</p>\n\n<p>Thanks\nHilal</p>",
      "rawMarkdown": "HI Julia:\n   Not a good time to change the UI on us this morning!.  We had a commit button , now it's save a new version, I wasn't able to tell what this is! .\n\nThanks\nHilal"
    },
    {
      "id": 792880,
      "postDate": "2020-03-31T15:59:52.757Z",
      "content": "<p>Hey, if we make a submission before the deadline, say at 11:30pm UTC March 31st 2020, and it is not evaluated before the competition deadline, will it be still considered for private test set evaluation? Assuming, no manual selection is being made.  </p>",
      "rawMarkdown": "Hey, if we make a submission before the deadline, say at 11:30pm UTC March 31st 2020, and it is not evaluated before the competition deadline, will it be still considered for private test set evaluation? Assuming, no manual selection is being made.  "
    },
    {
      "id": 792680,
      "postDate": "2020-03-31T13:09:06.240Z",
      "content": "<p>Excited to see the results! <a href=\"/juliaelliott\">@juliaelliott</a>, just to be sure, can we assume the new videos are in .mp4 format?</p>",
      "rawMarkdown": "Excited to see the results! @juliaelliott, just to be sure, can we assume the new videos are in .mp4 format?"
    },
    {
      "id": 791001,
      "postDate": "2020-03-30T03:08:28.020Z",
      "content": "<p><a href=\"/juliaelliott\">@juliaelliott</a> , Hi Julia, I find my notebook cannot stop and continue burning my GPU time, besides, I can not open another notebook because of maxinum GPU limit, when I click the notebook want to stop it , it turned out 404....Could you tell me how to solve this problem?</p>",
      "rawMarkdown": "@juliaelliott , Hi Julia, I find my notebook cannot stop and continue burning my GPU time, besides, I can not open another notebook because of maxinum GPU limit, when I click the notebook want to stop it , it turned out 404....Could you tell me how to solve this problem?"
    },
    {
      "id": 790930,
      "postDate": "2020-03-30T00:41:01.907Z",
      "content": "<p>How can we install matlab runtime in Kaggle kernel? </p>",
      "rawMarkdown": "How can we install matlab runtime in Kaggle kernel? "
    },
    {
      "id": 790772,
      "postDate": "2020-03-29T21:04:15.833Z",
      "content": "<p>Hi Julie,</p>\n\n<p>Is there any way to submit MATLAB scripts instead of Python scripts and our prediction results? Our validation accuracy is about 80% and it outperforms several popular deep learning models in our experiments. It is very hard for us to convert our matlab scripts to Python scripts in Kaggle Kernel.</p>\n\n<p>Thank you!</p>",
      "rawMarkdown": "Hi Julie,\n\nIs there any way to submit MATLAB scripts instead of Python scripts and our prediction results? Our validation accuracy is about 80% and it outperforms several popular deep learning models in our experiments. It is very hard for us to convert our matlab scripts to Python scripts in Kaggle Kernel.\n\nThank you!"
    },
    {
      "id": 787670,
      "postDate": "2020-03-27T01:29:04.210Z",
      "content": "<p>Hi Julia, Never received the GCP link for training data. Please help (email: annmargaret.tutu@icloud.com) </p>",
      "rawMarkdown": "Hi Julia, Never received the GCP link for training data. Please help (email: annmargaret.tutu@icloud.com) "
    },
    {
      "id": 792609,
      "postDate": "2020-03-31T11:21:57.397Z",
      "rawMarkdown": "",
      "isDeleted": true
    }
  ],
  "comments": [
    {
      "id": 788269,
      "author_name": "Arun",
      "author_url": "",
      "post_date": "2020-03-27T14:28:02.733000",
      "content": "<p>Hi Julia,\n I was also affected by the covid and could not work for about 4-5 weeks being in a \"hotspot\" location. It would be great to get a few weeks of extension if this is agreeable with the others. Thanks <a href=\"/juliaelliott\">@juliaelliott</a> </p>\n\n<p>Regards.</p>",
      "votes": 4,
      "replies": []
    },
    {
      "id": 787754,
      "author_name": "Hilal Shaath",
      "author_url": "",
      "post_date": "2020-03-27T03:47:56.030000",
      "content": "<p>I would like to ask for a 2 week extension. We have a pandemic that most likely affected the performance of many of us. We need additional time. If possible, can the dead line be extended. It's quite tight for many of us as it is.</p>",
      "votes": 1,
      "replies": []
    },
    {
      "id": 792565,
      "author_name": "adamRR",
      "author_url": "",
      "post_date": "2020-03-31T10:34:48.077000",
      "content": "",
      "votes": 1,
      "replies": [
        {
          "id": 792604,
          "author_name": "Sirish Somanchi",
          "author_url": "",
          "post_date": "2020-03-31T11:13:30.773000",
          "content": "<p><a href=\"/adamrr\">@adamrr</a>  A lot of us (115+) had been asking for Extension on the other thread (<a href=\"https://www.kaggle.com/c/deepfake-detection-challenge/discussion/135631\">https://www.kaggle.com/c/deepfake-detection-challenge/discussion/135631</a>), but all our requests were denied 😪 </p>",
          "votes": 1,
          "replies": []
        }
      ]
    },
    {
      "id": 791054,
      "author_name": "Brad Davis",
      "author_url": "",
      "post_date": "2020-03-30T04:33:11.640000",
      "content": "<p>Hi Julia - I've recently tried to make submissions but there were errors during the process - at the moment those submissions are counting towards my total allowed submissions per day and I am not able to try submitting again.  Based on my validation score I really think I have a chance of being competitive - I only have 4 submissions left given the timing towards the deadline and I'm quite nervous that those 4 submissions chances will result in errors as well and prevent me from making any submission at all.  Are there any workarounds regarding submission errors preventing me from trying to resubmit?  At the moment it's very difficult to troubleshoot the issue without \"risking it all.\"  Thanks in advance for your help.  </p>",
      "votes": 1,
      "replies": []
    },
    {
      "id": 793271,
      "author_name": "Yule Wang",
      "author_url": "",
      "post_date": "2020-03-31T23:18:53.577000",
      "content": "<p>Could you please give us several hours for selecting two submissions after UTC 23.59???? I know we must submit our last submission before 11.59 pm UTC. But the two submissions are still running, we are not able to see whether they are successfully submitted and see the scores. Also, we cannot select these two!</p>",
      "votes": 0,
      "replies": [
        {
          "id": 793289,
          "author_name": "Evgeny Kovalev",
          "author_url": "",
          "post_date": "2020-03-31T23:32:04.730000",
          "content": "<p>Please see this thread: <a href=\"https://www.kaggle.com/c/deepfake-detection-challenge/discussion/139986\">https://www.kaggle.com/c/deepfake-detection-challenge/discussion/139986</a></p>\n\n<p>It is a pity that you have such situation, but rules are equal for everyone. So if you wanted to make submit in the last day, you should have left maximum 9 hours for it till the deadline. If you didn't, then it's your fault and lesson for the next code competitions, sorry)</p>",
          "votes": 0,
          "replies": []
        }
      ]
    },
    {
      "id": 793109,
      "author_name": "Hilal Shaath",
      "author_url": "",
      "post_date": "2020-03-31T20:00:28.177000",
      "content": "<p>There seems to be a lot of submission error. The Notebook system is out of sync with data.  Some submissions look impossible!.</p>\n\n<p>Really not a good time to do those changes in the final day of this competition. </p>\n\n<p>I think many will be affected by this.</p>",
      "votes": 0,
      "replies": []
    },
    {
      "id": 793015,
      "author_name": "Hilal Shaath",
      "author_url": "",
      "post_date": "2020-03-31T18:14:25.867000",
      "content": "<p>HI Julia:\n   Not a good time to change the UI on us this morning!.  We had a commit button , now it's save a new version, I wasn't able to tell what this is! .</p>\n\n<p>Thanks\nHilal</p>",
      "votes": 0,
      "replies": []
    },
    {
      "id": 792880,
      "author_name": "Atulya Narayan",
      "author_url": "",
      "post_date": "2020-03-31T15:59:52.757000",
      "content": "<p>Hey, if we make a submission before the deadline, say at 11:30pm UTC March 31st 2020, and it is not evaluated before the competition deadline, will it be still considered for private test set evaluation? Assuming, no manual selection is being made.  </p>",
      "votes": 0,
      "replies": []
    },
    {
      "id": 792680,
      "author_name": "Carlo",
      "author_url": "",
      "post_date": "2020-03-31T13:09:06.240000",
      "content": "<p>Excited to see the results! <a href=\"/juliaelliott\">@juliaelliott</a>, just to be sure, can we assume the new videos are in .mp4 format?</p>",
      "votes": 0,
      "replies": []
    },
    {
      "id": 791001,
      "author_name": "tomholland",
      "author_url": "",
      "post_date": "2020-03-30T03:08:28.020000",
      "content": "<p><a href=\"/juliaelliott\">@juliaelliott</a> , Hi Julia, I find my notebook cannot stop and continue burning my GPU time, besides, I can not open another notebook because of maxinum GPU limit, when I click the notebook want to stop it , it turned out 404....Could you tell me how to solve this problem?</p>",
      "votes": 0,
      "replies": []
    },
    {
      "id": 790930,
      "author_name": "original daze",
      "author_url": "",
      "post_date": "2020-03-30T00:41:01.907000",
      "content": "<p>How can we install matlab runtime in Kaggle kernel? </p>",
      "votes": 0,
      "replies": []
    },
    {
      "id": 790772,
      "author_name": "original daze",
      "author_url": "",
      "post_date": "2020-03-29T21:04:15.833000",
      "content": "<p>Hi Julie,</p>\n\n<p>Is there any way to submit MATLAB scripts instead of Python scripts and our prediction results? Our validation accuracy is about 80% and it outperforms several popular deep learning models in our experiments. It is very hard for us to convert our matlab scripts to Python scripts in Kaggle Kernel.</p>\n\n<p>Thank you!</p>",
      "votes": 0,
      "replies": []
    },
    {
      "id": 787670,
      "author_name": "AnnMargaret Tutu",
      "author_url": "",
      "post_date": "2020-03-27T01:29:04.210000",
      "content": "<p>Hi Julia, Never received the GCP link for training data. Please help (email: annmargaret.tutu@icloud.com) </p>",
      "votes": 0,
      "replies": []
    },
    {
      "id": 792609,
      "author_name": "",
      "author_url": "",
      "post_date": "2020-03-31T11:21:57.397000",
      "content": "",
      "votes": 0,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "786293": "This is a reminder that the deadline for submission AND to select the 2 submissions that will be re-run by the host on the private test set is **March 31st 2020 at 11:59pm UTC**. To be clear, we are **only re-running selected submissions**. You will **not** receive a final private score on your other submissions. \n\nA few things to keep in mind (none of this should be new information):\n\n1. Ensure your code is responsive to a completely new test set and id's. Hard-coding of values will fail. The way this competition was designed -- with a hidden public leaderboard test set -- should have encouraged this to be the case, if you were able to submit and receive a public score without error.\n2. As stated in the [getting started](https://w.kaggle.com/c/deepfake-detection-challenge/overview/getting-started) page, the time constraints applied to the private test run will scale proportionally to the size of that test set.\n3. The file structure and naming of the files (`test_videos`, `sample_submission.csv`) currently available on the Data page and file paths expected within your notebook sessions will remain the same. The difference will be in the *contents* of those files (i.e. the videos themselves in test_videos will be swapped out for the actual private test videos).\n4. If you do not make a submission selections, the auto-selection will default to your highest public-leaderboard-scoring submission's notebook(s). This may or may not represent what you expect. So it's always safest to make the manual selections.\n5. If you are part of a team, only the team leader may make the final selections.\n6. Don't risk deleting your selected notebooks *or* associated external datasets before the final private leaderboard results are confirmed. If we're unable to access those notebooks/datasets during re-runs, you will not be able to receive a score on that notebook.\n\nBest of luck in the final days of this competition!",
    "788269": "Hi Julia,\n I was also affected by the covid and could not work for about 4-5 weeks being in a \"hotspot\" location. It would be great to get a few weeks of extension if this is agreeable with the others. Thanks @juliaelliott \n\nRegards.",
    "787754": "I would like to ask for a 2 week extension. We have a pandemic that most likely affected the performance of many of us. We need additional time. If possible, can the dead line be extended. It's quite tight for many of us as it is.\n",
    "792565": "",
    "791054": "Hi Julia - I've recently tried to make submissions but there were errors during the process - at the moment those submissions are counting towards my total allowed submissions per day and I am not able to try submitting again.  Based on my validation score I really think I have a chance of being competitive - I only have 4 submissions left given the timing towards the deadline and I'm quite nervous that those 4 submissions chances will result in errors as well and prevent me from making any submission at all.  Are there any workarounds regarding submission errors preventing me from trying to resubmit?  At the moment it's very difficult to troubleshoot the issue without \"risking it all.\"  Thanks in advance for your help.  ",
    "793271": "Could you please give us several hours for selecting two submissions after UTC 23.59???? I know we must submit our last submission before 11.59 pm UTC. But the two submissions are still running, we are not able to see whether they are successfully submitted and see the scores. Also, we cannot select these two!",
    "793109": "There seems to be a lot of submission error. The Notebook system is out of sync with data.  Some submissions look impossible!.\n\nReally not a good time to do those changes in the final day of this competition. \n\nI think many will be affected by this.",
    "793015": "HI Julia:\n   Not a good time to change the UI on us this morning!.  We had a commit button , now it's save a new version, I wasn't able to tell what this is! .\n\nThanks\nHilal",
    "792880": "Hey, if we make a submission before the deadline, say at 11:30pm UTC March 31st 2020, and it is not evaluated before the competition deadline, will it be still considered for private test set evaluation? Assuming, no manual selection is being made.  ",
    "792680": "Excited to see the results! @juliaelliott, just to be sure, can we assume the new videos are in .mp4 format?",
    "791001": "@juliaelliott , Hi Julia, I find my notebook cannot stop and continue burning my GPU time, besides, I can not open another notebook because of maxinum GPU limit, when I click the notebook want to stop it , it turned out 404....Could you tell me how to solve this problem?",
    "790930": "How can we install matlab runtime in Kaggle kernel? ",
    "790772": "Hi Julie,\n\nIs there any way to submit MATLAB scripts instead of Python scripts and our prediction results? Our validation accuracy is about 80% and it outperforms several popular deep learning models in our experiments. It is very hard for us to convert our matlab scripts to Python scripts in Kaggle Kernel.\n\nThank you!",
    "787670": "Hi Julia, Never received the GCP link for training data. Please help (email: annmargaret.tutu@icloud.com) ",
    "792609": ""
  }
}