As everyone is most likely aware, this competition is stretching the capabilities of our scoring system. Please use this thread to report details so we can investigate.
\n\nNote(s):
\n\nAs everyone is most likely aware, this competition is stretching the capabilities of our scoring system. Please use this thread to report details so we can investigate.
\n\nNote(s):
\n\nHi, \nI am using convertVideotoCSV.py but I am getting the error \"Evaluation Exception: Input string was not in a correct format\" when I submit my file. When I manually check my submission, it looks as if it conforms to the submission format. So I am really confused why this is happening? Also, this is a snippet from my submission: snippet_submission.csv
\n\nIf it is not too much to ask, it would be great if we could get detailed error messages(with possibly line number of input file). It is quite difficult to debug just knowing that there is an string which is not in correct format. Alternatively, making the evaluation script public could also help in debugging submission errors.
", "rawMarkdown": "Hi, \nI am using convertVideotoCSV.py but I am getting the error \"Evaluation Exception: Input string was not in a correct format\" when I submit my file. When I manually check my submission, it looks as if it conforms to the submission format. So I am really confused why this is happening? Also, this is a snippet from my submission: [snippet_submission.csv][1]\n\nIf it is not too much to ask, it would be great if we could get detailed error messages(with possibly line number of input file). It is quite difficult to debug just knowing that there is an string which is not in correct format. Alternatively, making the evaluation script public could also help in debugging submission errors.\n\n [1]: https://docs.google.com/spreadsheets/d/1ZTv5B8OzqM3DiBvS0BGj4xvdQjVxDUc5or2WrQ78Xlk/edit", "votes": null }, { "id": "326593", "postDate": "05/10/2018 01:44:19", "content": "my submission file is over 700Mb, and is over 300Mb after compression...for me , i have got the same error:Evaluation Exception: Sequence contains no elements. many times.\ncould you please check my submissions and tell me why i always get this error ? it's the problem of submission formate or something else?
", "rawMarkdown": "my submission file is over 700Mb, and is over 300Mb after compression...for me , i have got the same error:Evaluation Exception: Sequence contains no elements. many times.\ncould you please check my submissions and tell me why i always get this error ? it's the problem of submission formate or something else?", "votes": null }, { "id": "327745", "postDate": "05/12/2018 10:47:21", "content": "Hi Akshay, I have encounter the same error: \"Evaluation Exception: Sequence contains no elements\"\nAnd here is the snippet of my .csv file: whats_wrong.csv\nThe RLE encoding looks fine and I used the code as belows:
\n\ndef rle_encoding(x):\n\"\"\" Run-length encoding based on\nhttps://www.kaggle.com/rakhlin/fast-run-length-encoding-python\n\"\"\"\nassert x.dtype == np.bool\ndots = np.where(x.flatten() == 1)[0]\nrun_lengths = []\nprev = -2\nfor b in dots:\n if b > prev + 1:\n run_lengths.append([b, 0])\n run_lengths[-1][1] += 1\n prev = b\nreturn '|'.join('{} {}'.format(*pair) for pair in run_lengths)\nIt has been very frustrating on my side and I have no idea what causes the error. Anyone can help? Many thanks!
", "rawMarkdown": "Hi Akshay, I have encounter the same error: \"Evaluation Exception: Sequence contains no elements\"\nAnd here is the snippet of my .csv file: [whats_wrong.csv][1]\nThe RLE encoding looks fine and I used the code as belows:\n\n def rle_encoding(x):\n \"\"\" Run-length encoding based on\n https://www.kaggle.com/rakhlin/fast-run-length-encoding-python\n \"\"\"\n assert x.dtype == np.bool\n dots = np.where(x.flatten() == 1)[0]\n run_lengths = []\n prev = -2\n for b in dots:\n if b > prev + 1:\n run_lengths.append([b, 0])\n run_lengths[-1][1] += 1\n prev = b\n return '|'.join('{} {}'.format(*pair) for pair in run_lengths)\n\nIt has been very frustrating on my side and I have no idea what causes the error. Anyone can help? Many thanks!\n\n [1]: https://drive.google.com/open?id=1MZhNwD8JwRKEKesKha6ka-dEaoh8bjE0", "votes": null }, { "id": "328809", "postDate": "05/15/2018 05:51:23", "content": "Hi, I am seeing this error while submission.
\n\nEvaluation Exception: Index was outside the bounds of the array.
\n\nAnyone has any idea on this? thanks!
", "rawMarkdown": "Hi, I am seeing this error while submission.\n\nEvaluation Exception: Index was outside the bounds of the array.\n\nAnyone has any idea on this? thanks!", "votes": null }, { "id": "329496", "postDate": "05/16/2018 15:35:20", "content": "I was also getting similar error. On my side the problem was that for the last entry in submission.csv countPixels was not the same as the number of pixels in encodingPixels. I am using the same script for the rle_encoding of binary mask. I succeeded to submit after removing the last entry in submission.csv, though I still didn't debug in my script why this happens only for the last entry.
", "rawMarkdown": "I was also getting similar error. On my side the problem was that for the last entry in submission.csv countPixels was not the same as the number of pixels in encodingPixels. I am using the same script for the rle_encoding of binary mask. I succeeded to submit after removing the last entry in submission.csv, though I still didn't debug in my script why this happens only for the last entry.", "votes": null }, { "id": "329500", "postDate": "05/16/2018 15:43:09", "content": "When I submit my prediction (about 30M), my screen freezes after some time with the following message: Scoring, 219 seconds remaining. When I go to the leaderboard I see that the submission is completed and the execution time is 558 seconds with a score. Is this behavior normal? Do you also get the same?
", "rawMarkdown": "When I submit my prediction (about 30M), my screen freezes after some time with the following message: Scoring, 219 seconds remaining. When I go to the leaderboard I see that the submission is completed and the execution time is 558 seconds with a score. Is this behavior normal? Do you also get the same?", "votes": null }, { "id": "329715", "postDate": "05/17/2018 02:48:34", "content": "Hey Natasa,\n Thank you for your reply. I am glad someone used the RLE code above and obtain some non-error result (as a matter of fact, I used \"convertVideotoCSV.py\" and check they render the same result. May I ask what do you mean by \"last entry\". The last row, or the last entry in very row in \"EncodedPixels\"?
\n\nRegardless, I used the following code to check my RLE, and they seems fine and no wrong message:
\n\nsub = pd.read_csv('submission.csv')\n\nprint('Total test image: %d' % len(sub))\nfor i in range(len(sub)):\n pc = sub.loc[i]['PixelCount']\n rle = sub.loc[i]['EncodedPixels']\n rle_pc = 0\n rle_split = rle.split('|')\n start_prev = -1\n for p in range(len(rle_split)-1):\n start = int(rle_split[p].split(' ')[0])\n length = int(rle_split[p].split(' ')[1])\n if start <= start_prev:\n print(\"Wrong encoding for: %s\" % sub.loc[i][\"ImageId\"])\n start_prev = start + length\n rle_pc += length\n\n if i % 1000 == 0:\n print(i)\n if rle_pc != pc:\n print(\"Wrong: %d\" % i)\nI meant, the last low in my submission file. Probably I have a bug in my script of how I used the rle_encoding().
", "rawMarkdown": "I meant, the last low in my submission file. Probably I have a bug in my script of how I used the rle_encoding().", "votes": null }, { "id": "329930", "postDate": "05/17/2018 15:19:02", "content": "Hey guys, after hours of visual examination (and frustration) with very cryptic error messages, I finally have a non-zero submission with correct format.\nI have provided the following sanity check for those who are also experiencing error messages but have no idea what's your with your submission. Hope it helps!
", "rawMarkdown": "Hey guys, after hours of visual examination (and frustration) with very cryptic error messages, I finally have a non-zero submission with correct format.\nI have provided the following sanity check for those who are also experiencing error messages but have no idea what's your with your submission. Hope it helps!", "votes": null }, { "id": "330077", "postDate": "05/18/2018 01:50:56", "content": "Thanks for the script. Do you believe that having the exact length of 2710*3384 is an issue? The script issued by kaggle seems to create plenty of those values
\n\n if start_prev >= 2710 * 3384:\n print(\"Wrong start_prev for: %s\" % sub.loc[i][\"ImageId\"])\nThat would simply indicate that the object occupies the very last pixel (that would be the lower right corner).
\n\nEDIT: It looks like at least in my case your script does not help. I still get Evaluation Exception: Sequence contains no elements after running my submission through it. Although I removed the name check, as I renamed the files.
Hey Michael, did you quote the wrong code? My code you are referring to should be: if start_prev >= 2710 * 3384 right.
\n\nThe lower right corner index should be W*H-1 instead of W*H ( it should be happen so rarely only the very last pixel serve as an instance component anyway.)
\n\nI am sorry that you still have such error messages. I know it could be very frustrating, hope you can solve it soon.
", "rawMarkdown": "Hey Michael, did you quote the wrong code? My code you are referring to should be: **if start_prev >= 2710 * 3384** right.\n\nThe lower right corner index should be W*H-1 instead of W*H ( it should be happen so rarely only the very last pixel serve as an instance component anyway.)\n\nI am sorry that you still have such error messages. I know it could be very frustrating, hope you can solve it soon.", "votes": null }, { "id": "330090", "postDate": "05/18/2018 02:57:26", "content": "No, I quoted straight out of your script you linked above. To me the assumption seems valid, as you mention that the last pixel has position W*H-1. However even removing the lines violating that assumption did not help, the error must be somewhere else.
", "rawMarkdown": "No, I quoted straight out of your script you linked above. To me the assumption seems valid, as you mention that the last pixel has position W*H-1. However even removing the lines violating that assumption did not help, the error must be somewhere else.", "votes": null }, { "id": "330228", "postDate": "05/18/2018 10:16:49", "content": "Michael, one way to debug the error could be that you submit only one row of your submission file, if it works then 1000 etc., until you find out where you have an error. \nFor example, I had one stupid bug in my script, I forgot f.close(), so not all that had to be written into a submission file was written, so my last row of the submission file was not complete.
", "rawMarkdown": "Michael, one way to debug the error could be that you submit only one row of your submission file, if it works then 1000 etc., until you find out where you have an error. \nFor example, I had one stupid bug in my script, I forgot f.close(), so not all that had to be written into a submission file was written, so my last row of the submission file was not complete.", "votes": null }, { "id": "330343", "postDate": "05/18/2018 16:09:01", "content": "Well, even with binary search which is O(log(n)) (unfortunately in base 2) that would take me approximately 14 trials with my 10'000 examples, in the worst case almost 5 days worth of submissions.
\n\nBut it seems like the admins have heard us and released some additional information :)
", "rawMarkdown": "Well, even with binary search which is O(log(n)) (unfortunately in base 2) that would take me approximately 14 trials with my 10'000 examples, in the worst case almost 5 days worth of submissions.\n\nBut it seems like the admins have heard us and released some additional information :)", "votes": null }, { "id": "332861", "postDate": "05/23/2018 23:30:01", "content": "I get the following error:
\n\nThe given key was not present in the dictionary.
", "rawMarkdown": "I get the following error:\n\nThe given key was not present in the dictionary.", "votes": null }, { "id": "333133", "postDate": "05/24/2018 13:06:12", "content": "I also have the error \"Evaluation Exception: Input string was not in a correct format\" while submission. The problem is still there if I run the released evaluation code. I checked the submission format is correct. I am really appreciated if anyone could share your solution to this problem. Thanks!
", "rawMarkdown": "I also have the error \"Evaluation Exception: Input string was not in a correct format\" while submission. The problem is still there if I run the released evaluation code. I checked the submission format is correct. I am really appreciated if anyone could share your solution to this problem. Thanks!", "votes": null }, { "id": "334267", "postDate": "05/26/2018 20:18:31", "content": "Apparently now it's my turn to struggle with \"Evaluation Exception: Sequence contains no elements\". Can you at please output the line where it occures? Have no idea how to debag this.
", "rawMarkdown": "Apparently now it's my turn to struggle with \"Evaluation Exception: Sequence contains no elements\". Can you at please output the line where it occures? Have no idea how to debag this.", "votes": null }, { "id": "334826", "postDate": "05/28/2018 14:13:44", "content": "One error I had is that in some rows pixelCount was equal 0 and there was no encodedPixels column. When I removed such rows, my submission was successful.
", "rawMarkdown": "One error I had is that in some rows pixelCount was equal 0 and there was no encodedPixels column. When I removed such rows, my submission was successful.", "votes": null }, { "id": "334900", "postDate": "05/28/2018 17:21:39", "content": "This was the first hypothesis I checked...
", "rawMarkdown": "This was the first hypothesis I checked...", "votes": null }, { "id": "335980", "postDate": "05/30/2018 20:17:29", "content": "Can someone help?\nI have the prediction in format (3384, 2710, 7), where each channel represents a class. I am stuck how to generate csv file.
", "rawMarkdown": "Can someone help?\nI have the prediction in format (3384, 2710, 7), where each channel represents a class. I am stuck how to generate csv file.", "votes": null } ], "comments": [ { "id": 326525, "author_name": "akshaytrivedi", "author_url": "", "post_date": "05/09/2018 21:37:13", "content": "Hi, \nI am using convertVideotoCSV.py but I am getting the error \"Evaluation Exception: Input string was not in a correct format\" when I submit my file. When I manually check my submission, it looks as if it conforms to the submission format. So I am really confused why this is happening? Also, this is a snippet from my submission: snippet_submission.csv
\n\nIf it is not too much to ask, it would be great if we could get detailed error messages(with possibly line number of input file). It is quite difficult to debug just knowing that there is an string which is not in correct format. Alternatively, making the evaluation script public could also help in debugging submission errors.
", "votes": null, "replies": [ { "id": 327745, "author_name": "stevenwudi", "author_url": "", "post_date": "05/12/2018 10:47:21", "content": "Hi Akshay, I have encounter the same error: \"Evaluation Exception: Sequence contains no elements\"\nAnd here is the snippet of my .csv file: whats_wrong.csv\nThe RLE encoding looks fine and I used the code as belows:
\n\ndef rle_encoding(x):\n\"\"\" Run-length encoding based on\nhttps://www.kaggle.com/rakhlin/fast-run-length-encoding-python\n\"\"\"\nassert x.dtype == np.bool\ndots = np.where(x.flatten() == 1)[0]\nrun_lengths = []\nprev = -2\nfor b in dots:\n if b > prev + 1:\n run_lengths.append([b, 0])\n run_lengths[-1][1] += 1\n prev = b\nreturn '|'.join('{} {}'.format(*pair) for pair in run_lengths)\nIt has been very frustrating on my side and I have no idea what causes the error. Anyone can help? Many thanks!
", "votes": null, "replies": [] }, { "id": 329496, "author_name": "natasa1", "author_url": "", "post_date": "05/16/2018 15:35:20", "content": "I was also getting similar error. On my side the problem was that for the last entry in submission.csv countPixels was not the same as the number of pixels in encodingPixels. I am using the same script for the rle_encoding of binary mask. I succeeded to submit after removing the last entry in submission.csv, though I still didn't debug in my script why this happens only for the last entry.
", "votes": null, "replies": [] }, { "id": 329715, "author_name": "stevenwudi", "author_url": "", "post_date": "05/17/2018 02:48:34", "content": "Hey Natasa,\n Thank you for your reply. I am glad someone used the RLE code above and obtain some non-error result (as a matter of fact, I used \"convertVideotoCSV.py\" and check they render the same result. May I ask what do you mean by \"last entry\". The last row, or the last entry in very row in \"EncodedPixels\"?
\n\nRegardless, I used the following code to check my RLE, and they seems fine and no wrong message:
\n\nsub = pd.read_csv('submission.csv')\n\nprint('Total test image: %d' % len(sub))\nfor i in range(len(sub)):\n pc = sub.loc[i]['PixelCount']\n rle = sub.loc[i]['EncodedPixels']\n rle_pc = 0\n rle_split = rle.split('|')\n start_prev = -1\n for p in range(len(rle_split)-1):\n start = int(rle_split[p].split(' ')[0])\n length = int(rle_split[p].split(' ')[1])\n if start <= start_prev:\n print(\"Wrong encoding for: %s\" % sub.loc[i][\"ImageId\"])\n start_prev = start + length\n rle_pc += length\n\n if i % 1000 == 0:\n print(i)\n if rle_pc != pc:\n print(\"Wrong: %d\" % i)\nI meant, the last low in my submission file. Probably I have a bug in my script of how I used the rle_encoding().
", "votes": null, "replies": [] }, { "id": 333133, "author_name": "lvying603", "author_url": "", "post_date": "05/24/2018 13:06:12", "content": "I also have the error \"Evaluation Exception: Input string was not in a correct format\" while submission. The problem is still there if I run the released evaluation code. I checked the submission format is correct. I am really appreciated if anyone could share your solution to this problem. Thanks!
", "votes": null, "replies": [] } ] }, { "id": 326593, "author_name": "yh673025667", "author_url": "", "post_date": "05/10/2018 01:44:19", "content": "my submission file is over 700Mb, and is over 300Mb after compression...for me , i have got the same error:Evaluation Exception: Sequence contains no elements. many times.\ncould you please check my submissions and tell me why i always get this error ? it's the problem of submission formate or something else?
", "votes": null, "replies": [] }, { "id": 328809, "author_name": "darylclimb", "author_url": "", "post_date": "05/15/2018 05:51:23", "content": "Hi, I am seeing this error while submission.
\n\nEvaluation Exception: Index was outside the bounds of the array.
\n\nAnyone has any idea on this? thanks!
", "votes": null, "replies": [] }, { "id": 329500, "author_name": "natasa1", "author_url": "", "post_date": "05/16/2018 15:43:09", "content": "When I submit my prediction (about 30M), my screen freezes after some time with the following message: Scoring, 219 seconds remaining. When I go to the leaderboard I see that the submission is completed and the execution time is 558 seconds with a score. Is this behavior normal? Do you also get the same?
", "votes": null, "replies": [] }, { "id": 329930, "author_name": "stevenwudi", "author_url": "", "post_date": "05/17/2018 15:19:02", "content": "Hey guys, after hours of visual examination (and frustration) with very cryptic error messages, I finally have a non-zero submission with correct format.\nI have provided the following sanity check for those who are also experiencing error messages but have no idea what's your with your submission. Hope it helps!
", "votes": null, "replies": [ { "id": 330077, "author_name": "michaelheinzer", "author_url": "", "post_date": "05/18/2018 01:50:56", "content": "Thanks for the script. Do you believe that having the exact length of 2710*3384 is an issue? The script issued by kaggle seems to create plenty of those values
\n\n if start_prev >= 2710 * 3384:\n print(\"Wrong start_prev for: %s\" % sub.loc[i][\"ImageId\"])\nThat would simply indicate that the object occupies the very last pixel (that would be the lower right corner).
\n\nEDIT: It looks like at least in my case your script does not help. I still get Evaluation Exception: Sequence contains no elements after running my submission through it. Although I removed the name check, as I renamed the files.
Hey Michael, did you quote the wrong code? My code you are referring to should be: if start_prev >= 2710 * 3384 right.
\n\nThe lower right corner index should be W*H-1 instead of W*H ( it should be happen so rarely only the very last pixel serve as an instance component anyway.)
\n\nI am sorry that you still have such error messages. I know it could be very frustrating, hope you can solve it soon.
", "votes": null, "replies": [] }, { "id": 330090, "author_name": "michaelheinzer", "author_url": "", "post_date": "05/18/2018 02:57:26", "content": "No, I quoted straight out of your script you linked above. To me the assumption seems valid, as you mention that the last pixel has position W*H-1. However even removing the lines violating that assumption did not help, the error must be somewhere else.
", "votes": null, "replies": [] }, { "id": 330228, "author_name": "natasa1", "author_url": "", "post_date": "05/18/2018 10:16:49", "content": "Michael, one way to debug the error could be that you submit only one row of your submission file, if it works then 1000 etc., until you find out where you have an error. \nFor example, I had one stupid bug in my script, I forgot f.close(), so not all that had to be written into a submission file was written, so my last row of the submission file was not complete.
", "votes": null, "replies": [] }, { "id": 330343, "author_name": "michaelheinzer", "author_url": "", "post_date": "05/18/2018 16:09:01", "content": "Well, even with binary search which is O(log(n)) (unfortunately in base 2) that would take me approximately 14 trials with my 10'000 examples, in the worst case almost 5 days worth of submissions.
\n\nBut it seems like the admins have heard us and released some additional information :)
", "votes": null, "replies": [] } ] }, { "id": 332861, "author_name": "dskswu", "author_url": "", "post_date": "05/23/2018 23:30:01", "content": "I get the following error:
\n\nThe given key was not present in the dictionary.
", "votes": null, "replies": [] }, { "id": 334267, "author_name": "azakhtyamov", "author_url": "", "post_date": "05/26/2018 20:18:31", "content": "Apparently now it's my turn to struggle with \"Evaluation Exception: Sequence contains no elements\". Can you at please output the line where it occures? Have no idea how to debag this.
", "votes": null, "replies": [ { "id": 334826, "author_name": "natasa1", "author_url": "", "post_date": "05/28/2018 14:13:44", "content": "One error I had is that in some rows pixelCount was equal 0 and there was no encodedPixels column. When I removed such rows, my submission was successful.
", "votes": null, "replies": [] }, { "id": 334900, "author_name": "azakhtyamov", "author_url": "", "post_date": "05/28/2018 17:21:39", "content": "This was the first hypothesis I checked...
", "votes": null, "replies": [] } ] }, { "id": 335980, "author_name": "waseemabas", "author_url": "", "post_date": "05/30/2018 20:17:29", "content": "Can someone help?\nI have the prediction in format (3384, 2710, 7), where each channel represents a class. I am stuck how to generate csv file.
", "votes": null, "replies": [] } ], "raw_markdown_by_id": { "326444": "As everyone is most likely aware, this competition is stretching the capabilities of our scoring system. Please use this thread to report details so we can investigate.\n\nNote(s):\n\n - It appears that larger submission files are causing some issues. If you're raw file is > 75 Mb, please compress before submission and try again. \n - If you had an error, and then the same file worked (or worked after compression), we'd like to know that also.", "326525": "Hi, \nI am using convertVideotoCSV.py but I am getting the error \"Evaluation Exception: Input string was not in a correct format\" when I submit my file. When I manually check my submission, it looks as if it conforms to the submission format. So I am really confused why this is happening? Also, this is a snippet from my submission: [snippet_submission.csv][1]\n\nIf it is not too much to ask, it would be great if we could get detailed error messages(with possibly line number of input file). It is quite difficult to debug just knowing that there is an string which is not in correct format. Alternatively, making the evaluation script public could also help in debugging submission errors.\n\n [1]: https://docs.google.com/spreadsheets/d/1ZTv5B8OzqM3DiBvS0BGj4xvdQjVxDUc5or2WrQ78Xlk/edit", "326593": "my submission file is over 700Mb, and is over 300Mb after compression...for me , i have got the same error:Evaluation Exception: Sequence contains no elements. many times.\ncould you please check my submissions and tell me why i always get this error ? it's the problem of submission formate or something else?", "327745": "Hi Akshay, I have encounter the same error: \"Evaluation Exception: Sequence contains no elements\"\nAnd here is the snippet of my .csv file: [whats_wrong.csv][1]\nThe RLE encoding looks fine and I used the code as belows:\n\n def rle_encoding(x):\n \"\"\" Run-length encoding based on\n https://www.kaggle.com/rakhlin/fast-run-length-encoding-python\n \"\"\"\n assert x.dtype == np.bool\n dots = np.where(x.flatten() == 1)[0]\n run_lengths = []\n prev = -2\n for b in dots:\n if b > prev + 1:\n run_lengths.append([b, 0])\n run_lengths[-1][1] += 1\n prev = b\n return '|'.join('{} {}'.format(*pair) for pair in run_lengths)\n\nIt has been very frustrating on my side and I have no idea what causes the error. Anyone can help? Many thanks!\n\n [1]: https://drive.google.com/open?id=1MZhNwD8JwRKEKesKha6ka-dEaoh8bjE0", "328809": "Hi, I am seeing this error while submission.\n\nEvaluation Exception: Index was outside the bounds of the array.\n\nAnyone has any idea on this? thanks!", "329496": "I was also getting similar error. On my side the problem was that for the last entry in submission.csv countPixels was not the same as the number of pixels in encodingPixels. I am using the same script for the rle_encoding of binary mask. I succeeded to submit after removing the last entry in submission.csv, though I still didn't debug in my script why this happens only for the last entry.", "329500": "When I submit my prediction (about 30M), my screen freezes after some time with the following message: Scoring, 219 seconds remaining. When I go to the leaderboard I see that the submission is completed and the execution time is 558 seconds with a score. Is this behavior normal? Do you also get the same?", "329715": "Hey Natasa,\n Thank you for your reply. I am glad someone used the RLE code above and obtain some non-error result (as a matter of fact, I used \"convertVideotoCSV.py\" and check they render the same result. May I ask what do you mean by \"last entry\". The last row, or the last entry in very row in \"EncodedPixels\"?\n \nRegardless, I used the following code to check my RLE, and they seems fine and no wrong message:\n\n sub = pd.read_csv('submission.csv')\n \n print('Total test image: %d' % len(sub))\n for i in range(len(sub)):\n pc = sub.loc[i]['PixelCount']\n rle = sub.loc[i]['EncodedPixels']\n rle_pc = 0\n rle_split = rle.split('|')\n start_prev = -1\n for p in range(len(rle_split)-1):\n start = int(rle_split[p].split(' ')[0])\n length = int(rle_split[p].split(' ')[1])\n if start <= start_prev:\n print(\"Wrong encoding for: %s\" % sub.loc[i][\"ImageId\"])\n start_prev = start + length\n rle_pc += length\n \n if i % 1000 == 0:\n print(i)\n if rle_pc != pc:\n print(\"Wrong: %d\" % i)", "329795": "I meant, the last low in my submission file. Probably I have a bug in my script of how I used the rle_encoding().", "329930": "Hey guys, after hours of visual examination (and frustration) with very cryptic error messages, I finally have a non-zero submission with correct format.\nI have provided the following sanity check for those who are also experiencing error messages but have no idea what's your with your submission. Hope it helps!", "330077": "Thanks for the script. Do you believe that having the exact length of 2710*3384 is an issue? The script issued by kaggle seems to create plenty of those values\n\n if start_prev >= 2710 * 3384:\n print(\"Wrong start_prev for: %s\" % sub.loc[i][\"ImageId\"])\n\nThat would simply indicate that the object occupies the very last pixel (that would be the lower right corner).\n\nEDIT: It looks like at least in my case your script does not help. I still get `Evaluation Exception: Sequence contains no elements` after running my submission through it. Although I removed the name check, as I renamed the files.", "330080": "Hey Michael, did you quote the wrong code? My code you are referring to should be: **if start_prev >= 2710 * 3384** right.\n\nThe lower right corner index should be W*H-1 instead of W*H ( it should be happen so rarely only the very last pixel serve as an instance component anyway.)\n\nI am sorry that you still have such error messages. I know it could be very frustrating, hope you can solve it soon.", "330090": "No, I quoted straight out of your script you linked above. To me the assumption seems valid, as you mention that the last pixel has position W*H-1. However even removing the lines violating that assumption did not help, the error must be somewhere else.", "330228": "Michael, one way to debug the error could be that you submit only one row of your submission file, if it works then 1000 etc., until you find out where you have an error. \nFor example, I had one stupid bug in my script, I forgot f.close(), so not all that had to be written into a submission file was written, so my last row of the submission file was not complete.", "330343": "Well, even with binary search which is O(log(n)) (unfortunately in base 2) that would take me approximately 14 trials with my 10'000 examples, in the worst case almost 5 days worth of submissions.\n\nBut it seems like the admins have heard us and released some additional information :)", "332861": "I get the following error:\n\nThe given key was not present in the dictionary.", "333133": "I also have the error \"Evaluation Exception: Input string was not in a correct format\" while submission. The problem is still there if I run the released evaluation code. I checked the submission format is correct. I am really appreciated if anyone could share your solution to this problem. Thanks!", "334267": "Apparently now it's my turn to struggle with \"Evaluation Exception: Sequence contains no elements\". Can you at please output the line where it occures? Have no idea how to debag this.", "334826": "One error I had is that in some rows pixelCount was equal 0 and there was no encodedPixels column. When I removed such rows, my submission was successful.", "334900": "This was the first hypothesis I checked...", "335980": "Can someone help?\nI have the prediction in format (3384, 2710, 7), where each channel represents a class. I am stuck how to generate csv file." }, "source": "meta" }