Has any one tried using anatomical structures segmention files for the private test submission data?
\nI`m trying to use for some filtering but the submission fails with 'Submission Scoring Error' even though it runs fine with the public test data.
\nI checked that i don`t have anything hardcoded and everything looks fine.
\nI also submitted a solution that does use the file but just tries to open and it fails after a while (i think one of the slices might either not have that json or the json is corrupted somehow).
\nHas anyone else run into this problem?
", "messages": [ { "id": "1170702", "postDate": "01/26/2021 11:50:08", "content": "Has any one tried using anatomical structures segmention files for the private test submission data?
\nI`m trying to use for some filtering but the submission fails with 'Submission Scoring Error' even though it runs fine with the public test data.
\nI checked that i don`t have anything hardcoded and everything looks fine.
\nI also submitted a solution that does use the file but just tries to open and it fails after a while (i think one of the slices might either not have that json or the json is corrupted somehow).
\nHas anyone else run into this problem?
", "rawMarkdown": "Has any one tried using anatomical structures segmention files for the private test submission data?\n\nI`m trying to use for some filtering but the submission fails with 'Submission Scoring Error' even though it runs fine with the public test data. \n\nI checked that i don`t have anything hardcoded and everything looks fine.\n\nI also submitted a solution that does use the file but just tries to open and it fails after a while (i think one of the slices might either not have that json or the json is corrupted somehow).\n\nHas anyone else run into this problem?", "votes": null }, { "id": "1171575", "postDate": "01/27/2021 00:40:55", "content": "Not used yet…
", "rawMarkdown": "Not used yet...", "votes": null }, { "id": "1175919", "postDate": "01/29/2021 12:06:29", "content": "We cannot know for sure, but you can catch an exception and turn off the filtering if something is wrong with the file. I am doing it this way, works just fine. But these annotations are not always precise.
", "rawMarkdown": "We cannot know for sure, but you can catch an exception and turn off the filtering if something is wrong with the file. I am doing it this way, works just fine. But these annotations are not always precise.", "votes": null }, { "id": "1176395", "postDate": "01/29/2021 15:53:51", "content": "Yeah, i managed to make it work, added exceptions.
\nI tried to speed up inference by rejecting frames that fall outside of the cortex but the map is not quite accurate and i end up missing some gloms (FN).
\nWe can see this in the train tiffs as well, some gloms fall outside of the cortex annot.
", "rawMarkdown": "Yeah, i managed to make it work, added exceptions. \n\nI tried to speed up inference by rejecting frames that fall outside of the cortex but the map is not quite accurate and i end up missing some gloms (FN). \n \nWe can see this in the train tiffs as well, some gloms fall outside of the cortex annot.", "votes": null }, { "id": "1261917", "postDate": "04/03/2021 15:12:53", "content": "Hi
\n@rosuluc have you been able to solve your issue? Because I am stuck on this very problem that when I use anatomical structures JSON for only performing segmentation for a certain region on Private Set I get an error of Submission Scoring Error. However, I don't get any errors on Public Set i.e I am able to generate submission file..
Can you please explain what you did to fix the issue? I am currently facing the same problem.
", "rawMarkdown": "Can you please explain what you did to fix the issue? I am currently facing the same problem.", "votes": null }, { "id": "1264353", "postDate": "04/06/2021 05:15:12", "content": "Based on the Data page under The Dataset, it would seem that the private test set does not have json files for anatomical segmentation:
\n\"Both the training and public test sets also include anatomical structure segmentations. They are intended to help you identify the various parts of the tissue.\"
HuBMAP-20-dataset_information.csv is probably not there for private test either.
\nHaving 2 sample_submission.csv files, one under train, one under test is another confusion.
So which sample_submission.csv file are you using the one inside the test images folder or the one in the root of the directory?
", "rawMarkdown": "So which sample_submission.csv file are you using the one inside the test images folder or the one in the root of the directory?", "votes": null }, { "id": "1264472", "postDate": "04/06/2021 07:30:45", "content": "For now using the one in the competition directory, not in test. Usually that is where the sample submission file exists.
", "rawMarkdown": "For now using the one in the competition directory, not in test. Usually that is where the sample submission file exists.", "votes": null }, { "id": "1264507", "postDate": "04/06/2021 07:56:21", "content": "Is it possible to have a definite answer from the organizers! Because otherwise it doesn't worth investing time on speeding up the pipeline using anatomical structures!
", "rawMarkdown": "Is it possible to have a definite answer from the organizers! Because otherwise it doesn't worth investing time on speeding up the pipeline using anatomical structures!", "votes": null }, { "id": "1264520", "postDate": "04/06/2021 08:08:48", "content": "Hey, I gave up using the anatomical annotations a while back. In train data they are not very precise (some glomes fall outside of them) and my inference is quick (~30mins) and don`t need a speed-up by filtering the slides now.
", "rawMarkdown": "Hey, I gave up using the anatomical annotations a while back. In train data they are not very precise (some glomes fall outside of them) and my inference is quick (~30mins) and don`t need a speed-up by filtering the slides now.", "votes": null }, { "id": "1264528", "postDate": "04/06/2021 08:19:15", "content": "ok and can you tell me what Inference times are you getting?
", "rawMarkdown": "ok and can you tell me what Inference times are you getting?", "votes": null } ], "comments": [ { "id": 1171575, "author_name": "drtausamaru", "author_url": "", "post_date": "01/27/2021 00:40:55", "content": "Not used yet…
", "votes": null, "replies": [] }, { "id": 1175919, "author_name": "morituri", "author_url": "", "post_date": "01/29/2021 12:06:29", "content": "We cannot know for sure, but you can catch an exception and turn off the filtering if something is wrong with the file. I am doing it this way, works just fine. But these annotations are not always precise.
", "votes": null, "replies": [ { "id": 1176395, "author_name": "rosuluc", "author_url": "", "post_date": "01/29/2021 15:53:51", "content": "Yeah, i managed to make it work, added exceptions.
\nI tried to speed up inference by rejecting frames that fall outside of the cortex but the map is not quite accurate and i end up missing some gloms (FN).
\nWe can see this in the train tiffs as well, some gloms fall outside of the cortex annot.
", "votes": null, "replies": [] }, { "id": 1261921, "author_name": "nauyan", "author_url": "", "post_date": "04/03/2021 15:16:32", "content": "Can you please explain what you did to fix the issue? I am currently facing the same problem.
", "votes": null, "replies": [] } ] }, { "id": 1261917, "author_name": "nauyan", "author_url": "", "post_date": "04/03/2021 15:12:53", "content": "Hi
\n@rosuluc have you been able to solve your issue? Because I am stuck on this very problem that when I use anatomical structures JSON for only performing segmentation for a certain region on Private Set I get an error of Submission Scoring Error. However, I don't get any errors on Public Set i.e I am able to generate submission file..
Based on the Data page under The Dataset, it would seem that the private test set does not have json files for anatomical segmentation:
\n\"Both the training and public test sets also include anatomical structure segmentations. They are intended to help you identify the various parts of the tissue.\"
HuBMAP-20-dataset_information.csv is probably not there for private test either.
\nHaving 2 sample_submission.csv files, one under train, one under test is another confusion.
So which sample_submission.csv file are you using the one inside the test images folder or the one in the root of the directory?
", "votes": null, "replies": [] }, { "id": 1264472, "author_name": "something4kag", "author_url": "", "post_date": "04/06/2021 07:30:45", "content": "For now using the one in the competition directory, not in test. Usually that is where the sample submission file exists.
", "votes": null, "replies": [] } ] }, { "id": 1264507, "author_name": "voglinio", "author_url": "", "post_date": "04/06/2021 07:56:21", "content": "Is it possible to have a definite answer from the organizers! Because otherwise it doesn't worth investing time on speeding up the pipeline using anatomical structures!
", "votes": null, "replies": [] }, { "id": 1264520, "author_name": "rosuluc", "author_url": "", "post_date": "04/06/2021 08:08:48", "content": "Hey, I gave up using the anatomical annotations a while back. In train data they are not very precise (some glomes fall outside of them) and my inference is quick (~30mins) and don`t need a speed-up by filtering the slides now.
", "votes": null, "replies": [ { "id": 1264528, "author_name": "nauyan", "author_url": "", "post_date": "04/06/2021 08:19:15", "content": "ok and can you tell me what Inference times are you getting?
", "votes": null, "replies": [] } ] } ], "raw_markdown_by_id": { "1170702": "Has any one tried using anatomical structures segmention files for the private test submission data?\n\nI`m trying to use for some filtering but the submission fails with 'Submission Scoring Error' even though it runs fine with the public test data. \n\nI checked that i don`t have anything hardcoded and everything looks fine.\n\nI also submitted a solution that does use the file but just tries to open and it fails after a while (i think one of the slices might either not have that json or the json is corrupted somehow).\n\nHas anyone else run into this problem?", "1171575": "Not used yet...", "1175919": "We cannot know for sure, but you can catch an exception and turn off the filtering if something is wrong with the file. I am doing it this way, works just fine. But these annotations are not always precise.", "1176395": "Yeah, i managed to make it work, added exceptions. \n\nI tried to speed up inference by rejecting frames that fall outside of the cortex but the map is not quite accurate and i end up missing some gloms (FN). \n \nWe can see this in the train tiffs as well, some gloms fall outside of the cortex annot.", "1261917": "Hi\n@rosuluc have you been able to solve your issue? Because I am stuck on this very problem that when I use anatomical structures JSON for only performing segmentation for a certain region on Private Set I get an error of **Submission Scoring Error**. However, I don't get any errors on Public Set i.e I am able to generate submission file..", "1261921": "Can you please explain what you did to fix the issue? I am currently facing the same problem.", "1264353": "Based on the Data page under The Dataset, it would seem that the private test set does not have json files for anatomical segmentation: \n\"Both the training and public test sets also include anatomical structure segmentations. They are intended to help you identify the various parts of the tissue.\"\n\nHuBMAP-20-dataset_information.csv is probably not there for private test either.\nHaving 2 sample_submission.csv files, one under train, one under test is another confusion.", "1264461": "So which sample_submission.csv file are you using the one inside the test images folder or the one in the root of the directory?", "1264472": "For now using the one in the competition directory, not in test. Usually that is where the sample submission file exists.", "1264507": "Is it possible to have a definite answer from the organizers! Because otherwise it doesn't worth investing time on speeding up the pipeline using anatomical structures!", "1264520": "Hey, I gave up using the anatomical annotations a while back. In train data they are not very precise (some glomes fall outside of them) and my inference is quick (~30mins) and don`t need a speed-up by filtering the slides now.", "1264528": "ok and can you tell me what Inference times are you getting?" }, "source": "meta" }