This kernel is out of Resources when my submissions csv submit to kaggle.\nPlease tell me why this problem occurs.
\n\n【my kernel】\nhttps://www.kaggle.com/snakayama/vgg16-keras
", "messages": [ { "id": 569806, "postDate": "2019-07-07T11:37:47.253Z", "content": "This kernel is out of Resources when my submissions csv submit to kaggle.\nPlease tell me why this problem occurs.
\n\n【my kernel】\nhttps://www.kaggle.com/snakayama/vgg16-keras
", "rawMarkdown": "This kernel is out of Resources when my submissions csv submit to kaggle.\nPlease tell me why this problem occurs.\n\n【my kernel】\nhttps://www.kaggle.com/snakayama/vgg16-keras", "votes": 5 }, { "id": 570390, "postDate": "2019-07-08T08:05:40.447Z", "content": "Same problem!
", "rawMarkdown": "Same problem!", "votes": 1 }, { "id": 570088, "postDate": "2019-07-07T19:01:14.187Z", "content": "I have a same problem!
", "rawMarkdown": "I have a same problem! ", "votes": 1 }, { "id": 608178, "postDate": "2019-08-26T12:55:26.063Z", "content": "We are having the same issue with the Keras submission. Do we need to change something? The code takes like 30 minutes for the commit if you want some details. In the public score section, it took more than 5 hours for them to show us the Kernel out of resources error.
", "rawMarkdown": "We are having the same issue with the Keras submission. Do we need to change something? The code takes like 30 minutes for the commit if you want some details. In the public score section, it took more than 5 hours for them to show us the Kernel out of resources error. " }, { "id": 607616, "postDate": "2019-08-25T15:42:37.480Z", "content": "I have the same problem.
", "rawMarkdown": "I have the same problem." }, { "id": 593478, "postDate": "2019-08-06T16:49:56.317Z", "content": "I have encountered with same problem. Then I started to research about it like why this type of error occurs and how to solve it?
\nThis type of error may occur because we have different type of public test cases and private test cases, our memory may be bottle necked, if we have more number of images in private test case. Our model is not working properly.\nThis type of competitions are conducted to get more optimized solutions.\nSolution:- Make our model use less RAM then before, more optimized model, running in less time and easy to use.\nyou can use less epochs, best optimizer's and best preprocessing required.
Cursory glance at your kernel and it looks like you load all test images into memory for prediction. Remember that the hidden test data set has 30k+ images, so it's likely you are running out of RAM.
\n\nHave a look at using a data generator that reads batches from disk instead:
\n\nhttps://keras.io/preprocessing/image/
", "rawMarkdown": "Cursory glance at your kernel and it looks like you load all test images into memory for prediction. Remember that the hidden test data set has 30k+ images, so it's likely you are running out of RAM.\n\nHave a look at using a data generator that reads batches from disk instead:\n\nhttps://keras.io/preprocessing/image/" }, { "id": 581350, "postDate": "2019-07-21T19:44:11.607Z", "content": "I also have same problem. I've tried some of the solutions like dividing the epochs on fit generator in seperate cells and lowering the verbosity level. But it's still giving me the same error.
", "rawMarkdown": "I also have same problem. I've tried some of the solutions like dividing the epochs on fit generator in seperate cells and lowering the verbosity level. But it's still giving me the same error." }, { "id": 573865, "postDate": "2019-07-12T21:14:34.660Z", "content": "similar problem!
", "rawMarkdown": "similar problem!" }, { "id": 573608, "postDate": "2019-07-12T13:41:03.100Z", "content": "I am having the same problem. My kernel runs fine in the editor. Can someone please post the solution to this?
", "rawMarkdown": "I am having the same problem. My kernel runs fine in the editor. Can someone please post the solution to this?" }, { "id": 570415, "postDate": "2019-07-08T09:00:08.113Z", "content": "I have the same problem. What's wrong?
", "rawMarkdown": "I have the same problem. What's wrong?" }, { "id": 570117, "postDate": "2019-07-07T19:36:35.217Z", "content": "Same problem! It runs normally in the editor and when you commit it. The submission generates an error saying the kernel ran out of resources.
", "rawMarkdown": "Same problem! It runs normally in the editor and when you commit it. The submission generates an error saying the kernel ran out of resources." }, { "id": 570818, "postDate": "2019-07-08T19:49:34.887Z", "rawMarkdown": "", "isDeleted": true } ], "comments": [ { "id": 570390, "author_name": "Revathi Vijay", "author_url": "", "post_date": "2019-07-08T08:05:40.447000", "content": "Same problem!
", "votes": 1, "replies": [] }, { "id": 570088, "author_name": "E.Nik", "author_url": "", "post_date": "2019-07-07T19:01:14.187000", "content": "I have a same problem!
", "votes": 1, "replies": [] }, { "id": 608178, "author_name": "naitik", "author_url": "", "post_date": "2019-08-26T12:55:26.063000", "content": "We are having the same issue with the Keras submission. Do we need to change something? The code takes like 30 minutes for the commit if you want some details. In the public score section, it took more than 5 hours for them to show us the Kernel out of resources error.
", "votes": 0, "replies": [] }, { "id": 607616, "author_name": "palak sood", "author_url": "", "post_date": "2019-08-25T15:42:37.480000", "content": "I have the same problem.
", "votes": 0, "replies": [] }, { "id": 593478, "author_name": "Sainadh086", "author_url": "", "post_date": "2019-08-06T16:49:56.317000", "content": "I have encountered with same problem. Then I started to research about it like why this type of error occurs and how to solve it?
\nThis type of error may occur because we have different type of public test cases and private test cases, our memory may be bottle necked, if we have more number of images in private test case. Our model is not working properly.\nThis type of competitions are conducted to get more optimized solutions.\nSolution:- Make our model use less RAM then before, more optimized model, running in less time and easy to use.\nyou can use less epochs, best optimizer's and best preprocessing required.
Cursory glance at your kernel and it looks like you load all test images into memory for prediction. Remember that the hidden test data set has 30k+ images, so it's likely you are running out of RAM.
\n\nHave a look at using a data generator that reads batches from disk instead:
\n\nhttps://keras.io/preprocessing/image/
", "votes": 0, "replies": [] }, { "id": 581350, "author_name": "Alok Malik", "author_url": "", "post_date": "2019-07-21T19:44:11.607000", "content": "I also have same problem. I've tried some of the solutions like dividing the epochs on fit generator in seperate cells and lowering the verbosity level. But it's still giving me the same error.
", "votes": 0, "replies": [] }, { "id": 573865, "author_name": "Firat Gonen", "author_url": "", "post_date": "2019-07-12T21:14:34.660000", "content": "similar problem!
", "votes": 0, "replies": [] }, { "id": 573608, "author_name": "ranasac", "author_url": "", "post_date": "2019-07-12T13:41:03.100000", "content": "I am having the same problem. My kernel runs fine in the editor. Can someone please post the solution to this?
", "votes": 0, "replies": [] }, { "id": 570415, "author_name": "Bin Wang", "author_url": "", "post_date": "2019-07-08T09:00:08.113000", "content": "I have the same problem. What's wrong?
", "votes": 0, "replies": [] }, { "id": 570117, "author_name": "David Liang", "author_url": "", "post_date": "2019-07-07T19:36:35.217000", "content": "Same problem! It runs normally in the editor and when you commit it. The submission generates an error saying the kernel ran out of resources.
", "votes": 0, "replies": [] }, { "id": 570818, "author_name": "", "author_url": "", "post_date": "2019-07-08T19:49:34.887000", "content": "", "votes": 0, "replies": [] } ], "raw_markdown_by_id": { "569806": "This kernel is out of Resources when my submissions csv submit to kaggle.\nPlease tell me why this problem occurs.\n\n【my kernel】\nhttps://www.kaggle.com/snakayama/vgg16-keras", "570390": "Same problem!", "570088": "I have a same problem! ", "608178": "We are having the same issue with the Keras submission. Do we need to change something? The code takes like 30 minutes for the commit if you want some details. In the public score section, it took more than 5 hours for them to show us the Kernel out of resources error. ", "607616": "I have the same problem.", "593478": "I have encountered with same problem. Then I started to research about it like why this type of error occurs and how to solve it? \nThis type of error may occur because we have different type of public test cases and private test cases, our memory may be bottle necked, if we have more number of images in private test case. Our model is not working properly.\nThis type of competitions are conducted to get more optimized solutions.\nSolution:- Make our model use less RAM then before, more optimized model, running in less time and easy to use.\nyou can use less epochs, best optimizer's and best preprocessing required.", "581366": "Cursory glance at your kernel and it looks like you load all test images into memory for prediction. Remember that the hidden test data set has 30k+ images, so it's likely you are running out of RAM.\n\nHave a look at using a data generator that reads batches from disk instead:\n\nhttps://keras.io/preprocessing/image/", "581350": "I also have same problem. I've tried some of the solutions like dividing the epochs on fit generator in seperate cells and lowering the verbosity level. But it's still giving me the same error.", "573865": "similar problem!", "573608": "I am having the same problem. My kernel runs fine in the editor. Can someone please post the solution to this?", "570415": "I have the same problem. What's wrong?", "570117": "Same problem! It runs normally in the editor and when you commit it. The submission generates an error saying the kernel ran out of resources.", "570818": "" } }