Hello, I am new to the competition and I run some prediction on the 'stage1_test'.\nIt look okay when I visualize the masks in .png.\nHowever, when I submit the result, I get a score of 0.001.\nMy prediction is probably not in the top tier, but 0.001 seems too low for these predictions.
\n\nHere are my prediction in .png\nnew link: https://drive.google.com/file/d/143F4oEe-SE5c0TU7bd4TyNt-21WMbcpt/view?usp=sharing
\n\nand my code for saving .csv
\n\nf = open(\"submission.csv\", \"w\")\ncsv_writer = csv.writer(f)\ncsv_writer.writerow(['ImageId', 'EncodedPixels'])\nfor prediction in predictions:\n csv_writer.writerow([prediction_id, prediction_pixel])\nIf you have any suggestion, please feel free to help me out.
", "messages": [ { "id": "285724", "postDate": "02/20/2018 22:28:00", "content": "Hello, I am new to the competition and I run some prediction on the 'stage1_test'.\nIt look okay when I visualize the masks in .png.\nHowever, when I submit the result, I get a score of 0.001.\nMy prediction is probably not in the top tier, but 0.001 seems too low for these predictions.
\n\nHere are my prediction in .png\nnew link: https://drive.google.com/file/d/143F4oEe-SE5c0TU7bd4TyNt-21WMbcpt/view?usp=sharing
\n\nand my code for saving .csv
\n\nf = open(\"submission.csv\", \"w\")\ncsv_writer = csv.writer(f)\ncsv_writer.writerow(['ImageId', 'EncodedPixels'])\nfor prediction in predictions:\n csv_writer.writerow([prediction_id, prediction_pixel])\nIf you have any suggestion, please feel free to help me out.
", "rawMarkdown": "Hello, I am new to the competition and I run some prediction on the 'stage1_test'.\nIt look okay when I visualize the masks in .png.\nHowever, when I submit the result, I get a score of 0.001.\nMy prediction is probably not in the top tier, but 0.001 seems too low for these predictions.\n\nHere are my prediction in .png\nnew link: https://drive.google.com/file/d/143F4oEe-SE5c0TU7bd4TyNt-21WMbcpt/view?usp=sharing\n\nand my code for saving .csv\n\n f = open(\"submission.csv\", \"w\")\n csv_writer = csv.writer(f)\n csv_writer.writerow(['ImageId', 'EncodedPixels'])\n for prediction in predictions:\n csv_writer.writerow([prediction_id, prediction_pixel])\n\n\n\n\n\nIf you have any suggestion, please feel free to help me out.", "votes": null }, { "id": "285745", "postDate": "02/20/2018 23:23:19", "content": "SOLVED:\nThe problem was that I had submitted the prediction in one single RLE and the competition required one RLE per object in a image.
", "rawMarkdown": "SOLVED:\nThe problem was that I had submitted the prediction in one single RLE and the competition required one RLE per object in a image.", "votes": null }, { "id": "285748", "postDate": "02/20/2018 23:34:29", "content": "Does that mean that the RLE code in this kernel is wrong?
\n\nhttps://www.kaggle.com/kmader/nuclei-overview-to-submission
", "rawMarkdown": "Does that mean that the RLE code in this kernel is wrong?\n\nhttps://www.kaggle.com/kmader/nuclei-overview-to-submission", "votes": null }, { "id": "285749", "postDate": "02/20/2018 23:37:16", "content": "I'm not sure what you mean by this kernel, I have not published a kernel.
", "rawMarkdown": "I'm not sure what you mean by this kernel, I have not published a kernel.", "votes": null }, { "id": "285751", "postDate": "02/20/2018 23:39:20", "content": "I was referring to this kernel:
\n\nhttps://www.kaggle.com/kmader/nuclei-overview-to-submission
\n\nNot sure why the markdown link was not rendered properly
", "rawMarkdown": "I was referring to this kernel:\n\nhttps://www.kaggle.com/kmader/nuclei-overview-to-submission\n\nNot sure why the markdown link was not rendered properly", "votes": null }, { "id": "286224", "postDate": "02/21/2018 21:28:17", "content": "Nope I think the kernel explains it correctly since it's result in .csv has 1554+1 lines for 65 images.
", "rawMarkdown": "Nope I think the kernel explains it correctly since it's result in .csv has 1554+1 lines for 65 images.", "votes": null }, { "id": "286939", "postDate": "02/23/2018 03:50:53", "content": "Is 1554+1 line enough? i thought the competition expected We expect the solution file to have 4152 prediction rows. This file should have a header row.
Hi, I got a same problem as you had. But I did seperated those nucleus. My score still is 0.000. I can see that my predict is definitely not that bad. Do you have any idea what is happened? Thanks.
", "rawMarkdown": "Hi, I got a same problem as you had. But I did seperated those nucleus. My score still is 0.000. I can see that my predict is definitely not that bad. Do you have any idea what is happened? Thanks.", "votes": null }, { "id": "292504", "postDate": "03/08/2018 03:52:17", "content": "same here, https://www.kaggle.com/c/data-science-bowl-2018/discussion/50335
\n\nNot sure what to do!
", "rawMarkdown": "same here, https://www.kaggle.com/c/data-science-bowl-2018/discussion/50335\n\nNot sure what to do!", "votes": null } ], "comments": [ { "id": 285745, "author_name": "juheonjeon", "author_url": "", "post_date": "02/20/2018 23:23:19", "content": "SOLVED:\nThe problem was that I had submitted the prediction in one single RLE and the competition required one RLE per object in a image.
", "votes": null, "replies": [ { "id": 285748, "author_name": "mohdhasan", "author_url": "", "post_date": "02/20/2018 23:34:29", "content": "Does that mean that the RLE code in this kernel is wrong?
\n\nhttps://www.kaggle.com/kmader/nuclei-overview-to-submission
", "votes": null, "replies": [] }, { "id": 285749, "author_name": "juheonjeon", "author_url": "", "post_date": "02/20/2018 23:37:16", "content": "I'm not sure what you mean by this kernel, I have not published a kernel.
", "votes": null, "replies": [] }, { "id": 285751, "author_name": "mohdhasan", "author_url": "", "post_date": "02/20/2018 23:39:20", "content": "I was referring to this kernel:
\n\nhttps://www.kaggle.com/kmader/nuclei-overview-to-submission
\n\nNot sure why the markdown link was not rendered properly
", "votes": null, "replies": [] }, { "id": 286224, "author_name": "juheonjeon", "author_url": "", "post_date": "02/21/2018 21:28:17", "content": "Nope I think the kernel explains it correctly since it's result in .csv has 1554+1 lines for 65 images.
", "votes": null, "replies": [] }, { "id": 286939, "author_name": "decentmakeover", "author_url": "", "post_date": "02/23/2018 03:50:53", "content": "Is 1554+1 line enough? i thought the competition expected We expect the solution file to have 4152 prediction rows. This file should have a header row.
Hi, I got a same problem as you had. But I did seperated those nucleus. My score still is 0.000. I can see that my predict is definitely not that bad. Do you have any idea what is happened? Thanks.
", "votes": null, "replies": [ { "id": 292504, "author_name": "decentmakeover", "author_url": "", "post_date": "03/08/2018 03:52:17", "content": "same here, https://www.kaggle.com/c/data-science-bowl-2018/discussion/50335
\n\nNot sure what to do!
", "votes": null, "replies": [] } ] } ], "raw_markdown_by_id": { "285724": "Hello, I am new to the competition and I run some prediction on the 'stage1_test'.\nIt look okay when I visualize the masks in .png.\nHowever, when I submit the result, I get a score of 0.001.\nMy prediction is probably not in the top tier, but 0.001 seems too low for these predictions.\n\nHere are my prediction in .png\nnew link: https://drive.google.com/file/d/143F4oEe-SE5c0TU7bd4TyNt-21WMbcpt/view?usp=sharing\n\nand my code for saving .csv\n\n f = open(\"submission.csv\", \"w\")\n csv_writer = csv.writer(f)\n csv_writer.writerow(['ImageId', 'EncodedPixels'])\n for prediction in predictions:\n csv_writer.writerow([prediction_id, prediction_pixel])\n\n\n\n\n\nIf you have any suggestion, please feel free to help me out.", "285745": "SOLVED:\nThe problem was that I had submitted the prediction in one single RLE and the competition required one RLE per object in a image.", "285748": "Does that mean that the RLE code in this kernel is wrong?\n\nhttps://www.kaggle.com/kmader/nuclei-overview-to-submission", "285749": "I'm not sure what you mean by this kernel, I have not published a kernel.", "285751": "I was referring to this kernel:\n\nhttps://www.kaggle.com/kmader/nuclei-overview-to-submission\n\nNot sure why the markdown link was not rendered properly", "286224": "Nope I think the kernel explains it correctly since it's result in .csv has 1554+1 lines for 65 images.", "286939": "Is 1554+1 line enough? i thought the competition expected `We expect the solution file to have 4152 prediction rows. This file should have a header row.`", "291939": "Hi, I got a same problem as you had. But I did seperated those nucleus. My score still is 0.000. I can see that my predict is definitely not that bad. Do you have any idea what is happened? Thanks.", "292504": "same here, https://www.kaggle.com/c/data-science-bowl-2018/discussion/50335\n\nNot sure what to do!" }, "source": "meta" }