I am a beginner so I am not very well accustomed to working with kaggle notebooks. I have the following problem:
\n\nI am basically trying to augment images and storing them as a separate dataset so that my training kernel can directly use that dataset. I'm using 'plt.imsave' to save images while iterating through a loop. When I save and commit the kernel, the committed kernel only shows output visualizations, and no option to save the output contents as a dataset.
\n\nWhat am I doing wrong?
\n\nMy kernel: https://www.kaggle.com/dragonsan17/save-good-tiles-0/
\n\nForked from : https://www.kaggle.com/iafoss/panda-16x128x128-tiles
", "messages": [ { "id": 917493, "postDate": "2020-07-06T14:59:39.217Z", "content": "@dragonsan17 provide a snapshot of the code you are using otherwise we can't help
", "rawMarkdown": "@dragonsan17 provide a snapshot of the code you are using otherwise we can't help", "replies": [ { "id": 920362, "postDate": "2020-07-08T14:44:26.460Z", "content": "Hi. I have edited my post to include my kernel. Please check that
", "rawMarkdown": "Hi. I have edited my post to include my kernel. Please check that" }, { "id": 920397, "postDate": "2020-07-08T15:09:02.140Z", "content": "@dragonsan17 The problem is you directly save all the individual images without creating a folder first. Which is why the output shows individual images as a visualization instead of giving you the option to create a dataset. So first create a save directory and save the images. After committing you will be able to create a dataset.
", "rawMarkdown": "@dragonsan17 The problem is you directly save all the individual images without creating a folder first. Which is why the output shows individual images as a visualization instead of giving you the option to create a dataset. So first create a save directory and save the images. After committing you will be able to create a dataset.\n" }, { "id": 920523, "postDate": "2020-07-08T16:42:56.387Z", "content": "Tried that already. Look for my kernel now, i've made a fresh commit showing that. Only alternative working for me is making a zipfile. Is this how it is supposed to be?
", "rawMarkdown": "Tried that already. Look for my kernel now, i've made a fresh commit showing that. Only alternative working for me is making a zipfile. Is this how it is supposed to be?" }, { "id": 920529, "postDate": "2020-07-08T16:48:59.333Z", "content": "Oh yeah forgot to mention that part. The reason you need to create a zip or tar file of that save directory is because kaggle unzips that compressed file and creates the folder structure in the output. So yes, you will need to create a zip file of the save directory you create
", "rawMarkdown": "Oh yeah forgot to mention that part. The reason you need to create a zip or tar file of that save directory is because kaggle unzips that compressed file and creates the folder structure in the output. So yes, you will need to create a zip file of the save directory you create", "votes": 1 }, { "id": 920656, "postDate": "2020-07-08T18:26:17.700Z", "content": "Oh okay. Thanks man!
", "rawMarkdown": "Oh okay. Thanks man!" } ] }, { "id": 917465, "postDate": "2020-07-06T14:33:14.713Z", "content": "I am a beginner so I am not very well accustomed to working with kaggle notebooks. I have the following problem:
\n\nI am basically trying to augment images and storing them as a separate dataset so that my training kernel can directly use that dataset. I'm using 'plt.imsave' to save images while iterating through a loop. When I save and commit the kernel, the committed kernel only shows output visualizations, and no option to save the output contents as a dataset.
\n\nWhat am I doing wrong?
\n\nMy kernel: https://www.kaggle.com/dragonsan17/save-good-tiles-0/
\n\nForked from : https://www.kaggle.com/iafoss/panda-16x128x128-tiles
", "rawMarkdown": "I am a beginner so I am not very well accustomed to working with kaggle notebooks. I have the following problem:\n\nI am basically trying to augment images and storing them as a separate dataset so that my training kernel can directly use that dataset. I'm using 'plt.imsave' to save images while iterating through a loop. When I save and commit the kernel, the committed kernel only shows output visualizations, and no option to save the output contents as a dataset.\n\nWhat am I doing wrong?\n\nMy kernel: https://www.kaggle.com/dragonsan17/save-good-tiles-0/\n\nForked from : https://www.kaggle.com/iafoss/panda-16x128x128-tiles" } ], "comments": [ { "id": 917493, "author_name": "Yovin Yahathugoda", "author_url": "", "post_date": "2020-07-06T14:59:39.217000", "content": "@dragonsan17 provide a snapshot of the code you are using otherwise we can't help
", "votes": 0, "replies": [ { "id": 920362, "author_name": "M SANTOSH", "author_url": "", "post_date": "2020-07-08T14:44:26.460000", "content": "Hi. I have edited my post to include my kernel. Please check that
", "votes": 0, "replies": [] }, { "id": 920397, "author_name": "Yovin Yahathugoda", "author_url": "", "post_date": "2020-07-08T15:09:02.140000", "content": "@dragonsan17 The problem is you directly save all the individual images without creating a folder first. Which is why the output shows individual images as a visualization instead of giving you the option to create a dataset. So first create a save directory and save the images. After committing you will be able to create a dataset.
", "votes": 0, "replies": [] }, { "id": 920523, "author_name": "M SANTOSH", "author_url": "", "post_date": "2020-07-08T16:42:56.387000", "content": "Tried that already. Look for my kernel now, i've made a fresh commit showing that. Only alternative working for me is making a zipfile. Is this how it is supposed to be?
", "votes": 0, "replies": [] }, { "id": 920529, "author_name": "Yovin Yahathugoda", "author_url": "", "post_date": "2020-07-08T16:48:59.333000", "content": "Oh yeah forgot to mention that part. The reason you need to create a zip or tar file of that save directory is because kaggle unzips that compressed file and creates the folder structure in the output. So yes, you will need to create a zip file of the save directory you create
", "votes": 1, "replies": [] }, { "id": 920656, "author_name": "M SANTOSH", "author_url": "", "post_date": "2020-07-08T18:26:17.700000", "content": "Oh okay. Thanks man!
", "votes": 0, "replies": [] } ] } ], "raw_markdown_by_id": { "917493": "@dragonsan17 provide a snapshot of the code you are using otherwise we can't help", "917465": "I am a beginner so I am not very well accustomed to working with kaggle notebooks. I have the following problem:\n\nI am basically trying to augment images and storing them as a separate dataset so that my training kernel can directly use that dataset. I'm using 'plt.imsave' to save images while iterating through a loop. When I save and commit the kernel, the committed kernel only shows output visualizations, and no option to save the output contents as a dataset.\n\nWhat am I doing wrong?\n\nMy kernel: https://www.kaggle.com/dragonsan17/save-good-tiles-0/\n\nForked from : https://www.kaggle.com/iafoss/panda-16x128x128-tiles" } }