Welcome to the Panda prostate cancer classification challenge! The giant images used in digital pathology set up some interesting challenges that don't arise in standard image classification problems, so we're excited to see what you come up with.
\n\nThis also happens to be the largest competition dataset that we've been able to host through Kaggle's own datasets platform. This means that you can use TPUs to train your models within Kaggle notebooks, though you will still need to submit from a notebook that does not have a TPU attached. We're hopeful that the ability of TPUs to load very large patches of these images this will open up new approaches.
\n\nGood luck and happy Kaggling!
", "messages": [ { "id": 815568, "postDate": "2020-04-21T16:39:01.753Z", "content": "Welcome to the Panda prostate cancer classification challenge! The giant images used in digital pathology set up some interesting challenges that don't arise in standard image classification problems, so we're excited to see what you come up with.
\n\nThis also happens to be the largest competition dataset that we've been able to host through Kaggle's own datasets platform. This means that you can use TPUs to train your models within Kaggle notebooks, though you will still need to submit from a notebook that does not have a TPU attached. We're hopeful that the ability of TPUs to load very large patches of these images this will open up new approaches.
\n\nGood luck and happy Kaggling!
", "rawMarkdown": "Welcome to the Panda prostate cancer classification challenge! The giant images used in digital pathology set up some interesting challenges that don't arise in standard image classification problems, so we're excited to see what you come up with.\n\nThis also happens to be the largest competition dataset that we've been able to host through Kaggle's own datasets platform. This means that you can use TPUs to train your models within Kaggle notebooks, though you will still need to submit from a notebook that does not have a TPU attached. We're hopeful that the ability of TPUs to load very large patches of these images this will open up new approaches.\n\nGood luck and happy Kaggling!\n", "votes": 7 }, { "id": 898241, "postDate": "2020-06-23T11:52:17.667Z", "content": "@sohier \nI have 1 question,\nis it okay to pseudo label for test images at inference blind submission kernel?\nalthough it was previously discussed below (at wheat competition) and said to be OK by host, but i wanted to ask if it's okay here too.\nhttps://www.kaggle.com/c/global-wheat-detection/discussion/149148
", "rawMarkdown": "@sohier \nI have 1 question,\nis it okay to pseudo label for test images at inference blind submission kernel?\nalthough it was previously discussed below (at wheat competition) and said to be OK by host, but i wanted to ask if it's okay here too.\nhttps://www.kaggle.com/c/global-wheat-detection/discussion/149148", "votes": 1, "replies": [ { "id": 915732, "postDate": "2020-07-05T03:55:54.347Z", "content": "@sohier Silence means ok?\n@maggiemd @wouterbulten
", "rawMarkdown": "@sohier Silence means ok?\n@maggiemd @wouterbulten " }, { "id": 917501, "postDate": "2020-07-06T15:09:04.147Z", "content": "I can't think of any reason why it wouldn't be fine.
", "rawMarkdown": "I can't think of any reason why it wouldn't be fine." }, { "id": 917505, "postDate": "2020-07-06T15:14:07.490Z", "content": "Thanks! 😄
", "rawMarkdown": "Thanks! 😄 " } ] }, { "id": 815648, "postDate": "2020-04-21T17:46:19.337Z", "content": "Few questions about test data:\n1. The test.csv file contains only three rows is this expected (is remaining testset private)\n2. data page mentions test_images folder but currently there isn't one.\n3. The tiff files for rows in test.csv are not present in the dataset.
", "rawMarkdown": "Few questions about test data:\n1. The test.csv file contains only three rows is this expected (is remaining testset private)\n2. data page mentions test_images folder but currently there isn't one.\n3. The tiff files for rows in test.csv are not present in the dataset. \n", "votes": 1, "replies": [ { "id": 815674, "postDate": "2020-04-21T18:09:00.410Z", "content": "Good questions. The images in the test set are private, so they are only available when you submit your kernel. You can expect the test images to be in that folder and they are in the same format as the train images.
", "rawMarkdown": "Good questions. The images in the test set are private, so they are only available when you submit your kernel. You can expect the test images to be in that folder and they are in the same format as the train images.", "votes": 2 }, { "id": 815691, "postDate": "2020-04-21T18:34:58.160Z", "content": "In that case small testset with proper files and folders should be provided (like many other competitions). This helps to write inference code.
", "rawMarkdown": "In that case small testset with proper files and folders should be provided (like many other competitions). This helps to write inference code." }, { "id": 816575, "postDate": "2020-04-22T12:40:56.460Z", "content": "I'm using \"test_images\" folder when I submit my kernel and I'm still getting errors. What am I missing here? See commit #4, https://www.kaggle.com/yeayates21/panda-densenet-keras-starter
", "rawMarkdown": "I'm using \"test_images\" folder when I submit my kernel and I'm still getting errors. What am I missing here? See commit #4, https://www.kaggle.com/yeayates21/panda-densenet-keras-starter" }, { "id": 816606, "postDate": "2020-04-22T13:05:27.163Z", "content": "Have you already fixed your problem?
", "rawMarkdown": "Have you already fixed your problem?" }, { "id": 816985, "postDate": "2020-04-22T18:48:29.433Z", "content": "@maximblaze I haven't discovered a solution yet, and it appears that everyone is experiencing this issue. Given this, I'm hoping @wouterbulten can give us more information or a resolution.
", "rawMarkdown": "@maximblaze I haven't discovered a solution yet, and it appears that **everyone** is experiencing this issue. Given this, I'm hoping @wouterbulten can give us more information or a resolution." }, { "id": 816996, "postDate": "2020-04-22T18:57:48.723Z", "content": "We are aware that there are some problems regarding data access. The Kaggle team is looking into that as we speak. Note that it is expected that the downloadable test.csv only contains three rows; All test data is private and can only be accessed by the submitted notebooks.
\n\nComment by Maggie: https://kaggle.com/c/prostate-cancer-grade-assessment/discussion/145093#816800
", "rawMarkdown": "We are aware that there are some problems regarding data access. The Kaggle team is looking into that as we speak. Note that it is expected that the downloadable test.csv only contains three rows; All test data is private and can only be accessed by the submitted notebooks.\n\nComment by Maggie: https://kaggle.com/c/prostate-cancer-grade-assessment/discussion/145093#816800", "votes": 1 }, { "id": 823540, "postDate": "2020-04-27T18:03:05.267Z", "content": "@wouterbulten any updates on how people are supposed to make a submission?
", "rawMarkdown": "@wouterbulten any updates on how people are supposed to make a submission?", "votes": 1 } ] }, { "id": 815588, "postDate": "2020-04-21T16:55:30.390Z", "content": "\n\n\nThis also happens to be the largest competition dataset that we've been able to host through Kaggle's own datasets platform
\n
The data page shows (508 KB) please fix that 😄
", "rawMarkdown": "> This also happens to be the largest competition dataset that we've been able to host through Kaggle's own datasets platform\n\n\nThe data page shows (508 KB) please fix that 😄 ", "votes": 1, "replies": [ { "id": 815606, "postDate": "2020-04-21T17:12:04.753Z", "content": "You're looking at the figure for just train.csv.
Edit: I think I see which page you're referring to, we will fix that.
", "rawMarkdown": "You're looking at the figure for just `train.csv`.\n\nEdit: I think I see which page you're referring to, we will fix that.", "votes": 1 } ] }, { "id": 945216, "postDate": "2020-07-25T16:55:41.573Z", "content": "@sohier \nMy team was removed from public and private leader board.\nIssue on submitted notebook is appearing that 'notebook were submitted after competition date'.\nHow is it possible . My public rank was 41 and private 536. It has all disappeared now.\nCan someone please feedback. \nThanks.
", "rawMarkdown": "@sohier \nMy team was removed from public and private leader board.\nIssue on submitted notebook is appearing that 'notebook were submitted after competition date'.\nHow is it possible . My public rank was 41 and private 536. It has all disappeared now.\nCan someone please feedback. \nThanks.", "replies": [ { "id": 948053, "postDate": "2020-07-27T16:32:24.823Z", "content": "You are welcome to submit a compliance request here: https://www.kaggle.com/compliance
", "rawMarkdown": "You are welcome to submit a compliance request here: https://www.kaggle.com/compliance" } ] }, { "id": 882631, "postDate": "2020-06-12T02:26:32.990Z", "content": "!kaggle competitions download -c prostate-cancer-grade-assessment\nwhy i get this ??\n401 - Unauthorized
", "rawMarkdown": "!kaggle competitions download -c prostate-cancer-grade-assessment\nwhy i get this ??\n401 - Unauthorized\n\n" }, { "id": 880922, "postDate": "2020-06-10T16:37:06.167Z", "content": "I have a problem, I know that if we want to submit the submission.csv, we need to turn down the internet of notebook, but I need internet to download some packages before run, turn down the internet will lead error. How can I handle this problem.
", "rawMarkdown": "I have a problem, I know that if we want to submit the submission.csv, we need to turn down the internet of notebook, but I need internet to download some packages before run, turn down the internet will lead error. How can I handle this problem." }, { "id": 848583, "postDate": "2020-05-15T04:57:08.087Z", "content": "This discussion is quite old. But I also have the problem of an empty \"test_images\" folder. I assume that these are only visible when you commit a kernel, not before. But I tried doing a commit and still got this error. Thanks.
", "rawMarkdown": "This discussion is quite old. But I also have the problem of an empty \"test_images\" folder. I assume that these are only visible when you commit a kernel, not before. But I tried doing a commit and still got this error. Thanks. " }, { "id": 819723, "postDate": "2020-04-24T19:52:46.010Z", "content": "@sohier Like [train/test]_images, do we have test_label_masks folder for test set as well or not?
As there is no information about test_label_masks on data description page.
There are no test label masks.
", "rawMarkdown": "There are no test label masks.", "votes": 1 }, { "id": 819732, "postDate": "2020-04-24T19:59:47.773Z", "content": "Thanks for clarification.
", "rawMarkdown": "Thanks for clarification." } ] }, { "id": 816600, "postDate": "2020-04-22T13:02:31.980Z", "content": "Can anyone help me with understanding how can we make predictions?\nDataset does not include test images and labels… What shall we drop into submission file… \nAs I understood, we should train a model by predicting the ISUP score using train_csv and train images+labels and then we should normally make predictions using the built model on test images and test labels, am I wrong? So where can we find test images and labels for prediction?…. \nPlease help me!
", "rawMarkdown": "Can anyone help me with understanding how can we make predictions?\nDataset does not include test images and labels… What shall we drop into submission file… \nAs I understood, we should train a model by predicting the ISUP score using train_csv and train images+labels and then we should normally make predictions using the built model on test images and test labels, am I wrong? So where can we find test images and labels for prediction?…. \nPlease help me!", "replies": [ { "id": 862295, "postDate": "2020-05-26T14:18:11.233Z", "content": "Test dataset is not available for users, it's automatically being submitted when we train our model
", "rawMarkdown": "Test dataset is not available for users, it's automatically being submitted when we train our model" } ] }, { "id": 816152, "postDate": "2020-04-22T06:22:40.097Z", "content": "Unable to find test images, do we have mask_label_images for test set as well?
", "rawMarkdown": "Unable to find test images, do we have mask_label_images for test set as well?" }, { "id": 885024, "postDate": "2020-06-13T19:59:04.890Z", "rawMarkdown": "", "isDeleted": true }, { "id": 816002, "postDate": "2020-04-22T03:01:14.133Z", "rawMarkdown": "", "votes": -1, "isDeleted": true } ], "comments": [ { "id": 898241, "author_name": "Ryunosuke Ishizaki", "author_url": "", "post_date": "2020-06-23T11:52:17.667000", "content": "@sohier \nI have 1 question,\nis it okay to pseudo label for test images at inference blind submission kernel?\nalthough it was previously discussed below (at wheat competition) and said to be OK by host, but i wanted to ask if it's okay here too.\nhttps://www.kaggle.com/c/global-wheat-detection/discussion/149148
", "votes": 1, "replies": [ { "id": 915732, "author_name": "Ryunosuke Ishizaki", "author_url": "", "post_date": "2020-07-05T03:55:54.347000", "content": "@sohier Silence means ok?\n@maggiemd @wouterbulten
", "votes": 0, "replies": [] }, { "id": 917501, "author_name": "Sohier Dane", "author_url": "", "post_date": "2020-07-06T15:09:04.147000", "content": "I can't think of any reason why it wouldn't be fine.
", "votes": 0, "replies": [] }, { "id": 917505, "author_name": "Ryunosuke Ishizaki", "author_url": "", "post_date": "2020-07-06T15:14:07.490000", "content": "Thanks! 😄
", "votes": 0, "replies": [] } ] }, { "id": 815648, "author_name": "Dhananjay Raut", "author_url": "", "post_date": "2020-04-21T17:46:19.337000", "content": "Few questions about test data:\n1. The test.csv file contains only three rows is this expected (is remaining testset private)\n2. data page mentions test_images folder but currently there isn't one.\n3. The tiff files for rows in test.csv are not present in the dataset.
", "votes": 1, "replies": [ { "id": 815674, "author_name": "Wouter Bulten", "author_url": "", "post_date": "2020-04-21T18:09:00.410000", "content": "Good questions. The images in the test set are private, so they are only available when you submit your kernel. You can expect the test images to be in that folder and they are in the same format as the train images.
", "votes": 2, "replies": [] }, { "id": 815691, "author_name": "Dhananjay Raut", "author_url": "", "post_date": "2020-04-21T18:34:58.160000", "content": "In that case small testset with proper files and folders should be provided (like many other competitions). This helps to write inference code.
", "votes": 0, "replies": [] }, { "id": 816575, "author_name": "Matt Yates", "author_url": "", "post_date": "2020-04-22T12:40:56.460000", "content": "I'm using \"test_images\" folder when I submit my kernel and I'm still getting errors. What am I missing here? See commit #4, https://www.kaggle.com/yeayates21/panda-densenet-keras-starter
", "votes": 0, "replies": [] }, { "id": 816606, "author_name": "Maxim Doronkin", "author_url": "", "post_date": "2020-04-22T13:05:27.163000", "content": "Have you already fixed your problem?
", "votes": 0, "replies": [] }, { "id": 816985, "author_name": "Matt Yates", "author_url": "", "post_date": "2020-04-22T18:48:29.433000", "content": "@maximblaze I haven't discovered a solution yet, and it appears that everyone is experiencing this issue. Given this, I'm hoping @wouterbulten can give us more information or a resolution.
", "votes": 0, "replies": [] }, { "id": 816996, "author_name": "Wouter Bulten", "author_url": "", "post_date": "2020-04-22T18:57:48.723000", "content": "We are aware that there are some problems regarding data access. The Kaggle team is looking into that as we speak. Note that it is expected that the downloadable test.csv only contains three rows; All test data is private and can only be accessed by the submitted notebooks.
\n\nComment by Maggie: https://kaggle.com/c/prostate-cancer-grade-assessment/discussion/145093#816800
", "votes": 1, "replies": [] }, { "id": 823540, "author_name": "Matt Yates", "author_url": "", "post_date": "2020-04-27T18:03:05.267000", "content": "@wouterbulten any updates on how people are supposed to make a submission?
", "votes": 1, "replies": [] } ] }, { "id": 815588, "author_name": "Dhananjay Raut", "author_url": "", "post_date": "2020-04-21T16:55:30.390000", "content": "\n\n\nThis also happens to be the largest competition dataset that we've been able to host through Kaggle's own datasets platform
\n
The data page shows (508 KB) please fix that 😄
", "votes": 1, "replies": [ { "id": 815606, "author_name": "Sohier Dane", "author_url": "", "post_date": "2020-04-21T17:12:04.753000", "content": "You're looking at the figure for just train.csv.
Edit: I think I see which page you're referring to, we will fix that.
", "votes": 1, "replies": [] } ] }, { "id": 945216, "author_name": "Rajnish Chauhan", "author_url": "", "post_date": "2020-07-25T16:55:41.573000", "content": "@sohier \nMy team was removed from public and private leader board.\nIssue on submitted notebook is appearing that 'notebook were submitted after competition date'.\nHow is it possible . My public rank was 41 and private 536. It has all disappeared now.\nCan someone please feedback. \nThanks.
", "votes": 0, "replies": [ { "id": 948053, "author_name": "Sohier Dane", "author_url": "", "post_date": "2020-07-27T16:32:24.823000", "content": "You are welcome to submit a compliance request here: https://www.kaggle.com/compliance
", "votes": 0, "replies": [] } ] }, { "id": 882631, "author_name": "RoRonoA-TKO", "author_url": "", "post_date": "2020-06-12T02:26:32.990000", "content": "!kaggle competitions download -c prostate-cancer-grade-assessment\nwhy i get this ??\n401 - Unauthorized
", "votes": 0, "replies": [] }, { "id": 880922, "author_name": "carliu", "author_url": "", "post_date": "2020-06-10T16:37:06.167000", "content": "I have a problem, I know that if we want to submit the submission.csv, we need to turn down the internet of notebook, but I need internet to download some packages before run, turn down the internet will lead error. How can I handle this problem.
", "votes": 0, "replies": [] }, { "id": 848583, "author_name": "andy jennings", "author_url": "", "post_date": "2020-05-15T04:57:08.087000", "content": "This discussion is quite old. But I also have the problem of an empty \"test_images\" folder. I assume that these are only visible when you commit a kernel, not before. But I tried doing a commit and still got this error. Thanks.
", "votes": 0, "replies": [] }, { "id": 819723, "author_name": "Dracarys", "author_url": "", "post_date": "2020-04-24T19:52:46.010000", "content": "@sohier Like [train/test]_images, do we have test_label_masks folder for test set as well or not?
As there is no information about test_label_masks on data description page.
There are no test label masks.
", "votes": 1, "replies": [] }, { "id": 819732, "author_name": "Dracarys", "author_url": "", "post_date": "2020-04-24T19:59:47.773000", "content": "Thanks for clarification.
", "votes": 0, "replies": [] } ] }, { "id": 816600, "author_name": "Maxim Doronkin", "author_url": "", "post_date": "2020-04-22T13:02:31.980000", "content": "Can anyone help me with understanding how can we make predictions?\nDataset does not include test images and labels… What shall we drop into submission file… \nAs I understood, we should train a model by predicting the ISUP score using train_csv and train images+labels and then we should normally make predictions using the built model on test images and test labels, am I wrong? So where can we find test images and labels for prediction?…. \nPlease help me!
", "votes": 0, "replies": [ { "id": 862295, "author_name": "Kurian Benoy", "author_url": "", "post_date": "2020-05-26T14:18:11.233000", "content": "Test dataset is not available for users, it's automatically being submitted when we train our model
", "votes": 0, "replies": [] } ] }, { "id": 816152, "author_name": "Dracarys", "author_url": "", "post_date": "2020-04-22T06:22:40.097000", "content": "Unable to find test images, do we have mask_label_images for test set as well?
", "votes": 0, "replies": [] }, { "id": 885024, "author_name": "", "author_url": "", "post_date": "2020-06-13T19:59:04.890000", "content": "", "votes": 0, "replies": [] }, { "id": 816002, "author_name": "", "author_url": "", "post_date": "2020-04-22T03:01:14.133000", "content": "", "votes": -1, "replies": [] } ], "raw_markdown_by_id": { "815568": "Welcome to the Panda prostate cancer classification challenge! The giant images used in digital pathology set up some interesting challenges that don't arise in standard image classification problems, so we're excited to see what you come up with.\n\nThis also happens to be the largest competition dataset that we've been able to host through Kaggle's own datasets platform. This means that you can use TPUs to train your models within Kaggle notebooks, though you will still need to submit from a notebook that does not have a TPU attached. We're hopeful that the ability of TPUs to load very large patches of these images this will open up new approaches.\n\nGood luck and happy Kaggling!\n", "898241": "@sohier \nI have 1 question,\nis it okay to pseudo label for test images at inference blind submission kernel?\nalthough it was previously discussed below (at wheat competition) and said to be OK by host, but i wanted to ask if it's okay here too.\nhttps://www.kaggle.com/c/global-wheat-detection/discussion/149148", "815648": "Few questions about test data:\n1. The test.csv file contains only three rows is this expected (is remaining testset private)\n2. data page mentions test_images folder but currently there isn't one.\n3. The tiff files for rows in test.csv are not present in the dataset. \n", "815588": "> This also happens to be the largest competition dataset that we've been able to host through Kaggle's own datasets platform\n\n\nThe data page shows (508 KB) please fix that 😄 ", "945216": "@sohier \nMy team was removed from public and private leader board.\nIssue on submitted notebook is appearing that 'notebook were submitted after competition date'.\nHow is it possible . My public rank was 41 and private 536. It has all disappeared now.\nCan someone please feedback. \nThanks.", "882631": "!kaggle competitions download -c prostate-cancer-grade-assessment\nwhy i get this ??\n401 - Unauthorized\n\n", "880922": "I have a problem, I know that if we want to submit the submission.csv, we need to turn down the internet of notebook, but I need internet to download some packages before run, turn down the internet will lead error. How can I handle this problem.", "848583": "This discussion is quite old. But I also have the problem of an empty \"test_images\" folder. I assume that these are only visible when you commit a kernel, not before. But I tried doing a commit and still got this error. Thanks. ", "819723": "@sohier Like [train/test]_images, do we have `test_label_masks` folder for test set as well or not?\n\nAs there is no information about `test_label_masks` on data description page.", "816600": "Can anyone help me with understanding how can we make predictions?\nDataset does not include test images and labels… What shall we drop into submission file… \nAs I understood, we should train a model by predicting the ISUP score using train_csv and train images+labels and then we should normally make predictions using the built model on test images and test labels, am I wrong? So where can we find test images and labels for prediction?…. \nPlease help me!", "816152": "Unable to find test images, do we have mask_label_images for test set as well?", "885024": "", "816002": "" } }