![\"\"](\"https://www.googleapis.com/download/storage/v1/b/kaggle-user-content/o/inbox%2F654629%2F052c743f7de0ab6465b473cc7bdb026c%2F2019-06-29%2020_13_43-APTOS%202019%20Blindness%20Detection%20_%20Kaggle.png?generation=1561828650811900&alt=media\")
Dear All,\nWell, I have my own reservations regarding this new \"kernel only\" competition ,but let us put that aside for a second. I uploaded my model which was created externally,e.g. not in the kernel itself, and attached it to the kernel. Subsequently, I followed the instructions for submitting a new CSV, however, after **waiting for hours **for the rendering/submission to complete, I keep receiving errors without any further explanation and/or details regraring the source of the error.
\n\n
Dear All,\nWell, I have my own reservations regarding this new \"kernel only\" competition ,but let us put that aside for a second. I uploaded my model which was created externally,e.g. not in the kernel itself, and attached it to the kernel. Subsequently, I followed the instructions for submitting a new CSV, however, after **waiting for hours **for the rendering/submission to complete, I keep receiving errors without any further explanation and/or details regraring the source of the error.
\n\n
2 of them are out of time limit.... 9 hours is the maximum time you can have.
\n\nmake sure you dont have any errors and ur kernel runs on 6 times the data perfectly? are u using large batch size? are you using large images that may explode memory? these are some things you can check
", "rawMarkdown": "2 of them are out of time limit.... 9 hours is the maximum time you can have.\n\nmake sure you dont have any errors and ur kernel runs on 6 times the data perfectly? are u using large batch size? are you using large images that may explode memory? these are some things you can check", "votes": null }, { "id": "564602", "postDate": "06/29/2019 17:24:32", "content": "On Kernel Requirements says:
\n\n\"In synchronous KO competitions, note that a submission that results in an error--either within the kernel or within the process of scoring--will count against your daily submission limit and will not return the specific error message. This is necessary to prevent probing the private test set.
\n\nPlease see the Kernels-only FAQ for more information on how to submit.\"
", "rawMarkdown": "On Kernel Requirements says:\n\n\"In synchronous KO competitions, note that a submission that results in an error--either within the kernel or within the process of scoring--will count against your daily submission limit and will not return the specific error message. This is necessary to prevent probing the private test set.\n\nPlease see the Kernels-only FAQ for more information on how to submit.\"", "votes": null }, { "id": "564609", "postDate": "06/29/2019 17:37:00", "content": "The Kernel runs perfectly with a batch size of 1 on the default size of 224 and finishes in 10 minutes.
", "rawMarkdown": "The Kernel runs perfectly with a batch size of 1 on the default size of 224 and finishes in 10 minutes.", "votes": null }, { "id": "565303", "postDate": "06/30/2019 18:24:01", "content": "Same here. I preprocessed images after downloading on my computer and then uploaded them as dataset. Then trained and tested model based on those preprocessed images and compiled a submission.csv file. The kernel commits fine and makes a submission.csv. But the grader takes an hour and then says \"Error\" No additional details provided for this error.
\n\nWhat might be wrong ? @abhishek if you can see anything wrong in the above methodology please guide
", "rawMarkdown": "Same here. I preprocessed images after downloading on my computer and then uploaded them as dataset. Then trained and tested model based on those preprocessed images and compiled a submission.csv file. The kernel commits fine and makes a submission.csv. But the grader takes an hour and then says \"Error\" No additional details provided for this error. \n\nWhat might be wrong ? @abhishek if you can see anything wrong in the above methodology please guide", "votes": null }, { "id": "565404", "postDate": "06/30/2019 22:49:10", "content": "i just solved it as i realized index=False was not included when i was converting my file to csv. That resolved error. But now i am getting a 0.000 Public score :(
", "rawMarkdown": "i just solved it as i realized index=False was not included when i was converting my file to csv. That resolved error. But now i am getting a 0.000 Public score :(", "votes": null }, { "id": "565527", "postDate": "07/01/2019 04:46:16", "content": "Yes even I am also having the same issue if not doing the end to end process inside the kernel only.
", "rawMarkdown": "Yes even I am also having the same issue if not doing the end to end process inside the kernel only.", "votes": null }, { "id": "565529", "postDate": "07/01/2019 04:50:10", "content": "I face a similar issue. Even after getting an extremely good kappa score during training, My submissions return a 0 score after hours of time taken during submission. My overall kernel runtime is about 7800s roughly equivalent to 2 hours.
", "rawMarkdown": "I face a similar issue. Even after getting an extremely good kappa score during training, My submissions return a 0 score after hours of time taken during submission. My overall kernel runtime is about 7800s roughly equivalent to 2 hours.", "votes": null } ], "comments": [ { "id": 564600, "author_name": "abhishek", "author_url": "", "post_date": "06/29/2019 17:23:01", "content": "2 of them are out of time limit.... 9 hours is the maximum time you can have.
\n\nmake sure you dont have any errors and ur kernel runs on 6 times the data perfectly? are u using large batch size? are you using large images that may explode memory? these are some things you can check
", "votes": null, "replies": [ { "id": 564609, "author_name": "solomonk", "author_url": "", "post_date": "06/29/2019 17:37:00", "content": "The Kernel runs perfectly with a batch size of 1 on the default size of 224 and finishes in 10 minutes.
", "votes": null, "replies": [] } ] }, { "id": 564602, "author_name": "smrucv", "author_url": "", "post_date": "06/29/2019 17:24:32", "content": "On Kernel Requirements says:
\n\n\"In synchronous KO competitions, note that a submission that results in an error--either within the kernel or within the process of scoring--will count against your daily submission limit and will not return the specific error message. This is necessary to prevent probing the private test set.
\n\nPlease see the Kernels-only FAQ for more information on how to submit.\"
", "votes": null, "replies": [] }, { "id": 565303, "author_name": "shayanalibhatti", "author_url": "", "post_date": "06/30/2019 18:24:01", "content": "Same here. I preprocessed images after downloading on my computer and then uploaded them as dataset. Then trained and tested model based on those preprocessed images and compiled a submission.csv file. The kernel commits fine and makes a submission.csv. But the grader takes an hour and then says \"Error\" No additional details provided for this error.
\n\nWhat might be wrong ? @abhishek if you can see anything wrong in the above methodology please guide
", "votes": null, "replies": [ { "id": 565404, "author_name": "shayanalibhatti", "author_url": "", "post_date": "06/30/2019 22:49:10", "content": "i just solved it as i realized index=False was not included when i was converting my file to csv. That resolved error. But now i am getting a 0.000 Public score :(
", "votes": null, "replies": [] } ] }, { "id": 565527, "author_name": "ss1404", "author_url": "", "post_date": "07/01/2019 04:46:16", "content": "Yes even I am also having the same issue if not doing the end to end process inside the kernel only.
", "votes": null, "replies": [] }, { "id": 565529, "author_name": "thanatoz", "author_url": "", "post_date": "07/01/2019 04:50:10", "content": "I face a similar issue. Even after getting an extremely good kappa score during training, My submissions return a 0 score after hours of time taken during submission. My overall kernel runtime is about 7800s roughly equivalent to 2 hours.
", "votes": null, "replies": [] } ], "raw_markdown_by_id": { "564597": "Dear All,\nWell, I have my own reservations regarding this new \"kernel only\" competition ,but let us put that aside for a second. I uploaded my model which was created externally,e.g. not in the kernel itself, and attached it to the kernel. Subsequently, I followed the instructions for submitting a new CSV, however, after **waiting for hours **for the rendering/submission to complete, I keep receiving errors without any further explanation and/or details regraring the source of the error. \n\n", "564600": "2 of them are out of time limit.... 9 hours is the maximum time you can have.\n\nmake sure you dont have any errors and ur kernel runs on 6 times the data perfectly? are u using large batch size? are you using large images that may explode memory? these are some things you can check", "564602": "On Kernel Requirements says:\n\n\"In synchronous KO competitions, note that a submission that results in an error--either within the kernel or within the process of scoring--will count against your daily submission limit and will not return the specific error message. This is necessary to prevent probing the private test set.\n\nPlease see the Kernels-only FAQ for more information on how to submit.\"", "564609": "The Kernel runs perfectly with a batch size of 1 on the default size of 224 and finishes in 10 minutes.", "565303": "Same here. I preprocessed images after downloading on my computer and then uploaded them as dataset. Then trained and tested model based on those preprocessed images and compiled a submission.csv file. The kernel commits fine and makes a submission.csv. But the grader takes an hour and then says \"Error\" No additional details provided for this error. \n\nWhat might be wrong ? @abhishek if you can see anything wrong in the above methodology please guide", "565404": "i just solved it as i realized index=False was not included when i was converting my file to csv. That resolved error. But now i am getting a 0.000 Public score :(", "565527": "Yes even I am also having the same issue if not doing the end to end process inside the kernel only.", "565529": "I face a similar issue. Even after getting an extremely good kappa score during training, My submissions return a 0 score after hours of time taken during submission. My overall kernel runtime is about 7800s roughly equivalent to 2 hours." }, "source": "meta" }