{ "id": 400611, "title": "Test dataset prediction", "url": "/competitions/benetech-making-graphs-accessible/discussion/400611", "author_name": "", "post_date": "2023-04-09T10:56:31.903887300Z", "votes": 3, "comment_count": 5, "views": 0, "content": "

How do I predict results on the test dataset in my notebook?
\nMaybe its a bit silly question but I can not undetstand how to upload test images to my notebook for predictions.

", "messages": [ { "id": "2215498", "postDate": "04/09/2023 10:56:31", "content": "

How do I predict results on the test dataset in my notebook?
\nMaybe its a bit silly question but I can not undetstand how to upload test images to my notebook for predictions.

", "rawMarkdown": "How do I predict results on the test dataset in my notebook?\nMaybe its a bit silly question but I can not undetstand how to upload test images to my notebook for predictions.", "votes": null }, { "id": "2218495", "postDate": "04/11/2023 18:39:01", "content": "

When you submit your notebook for scoring, it will make predictions on the real test images, which are different from the ones you see when you edit the notebook.
\nSo first, your notebook must make valid predictions on the sample images.
\nSecond, you need to make sure you are making valid predictions.
\nI found that this was hard to achieve because my notebook threw exceptions when I submitted it. :(
\nBut I am still trying.
\nThe test folder contains 4000 per the description of the competition.
\nI am still learning about code competition.
\nI hope this helps.

", "rawMarkdown": "When you submit your notebook for scoring, it will make predictions on the real test images, which are different from the ones you see when you edit the notebook. \nSo first, your notebook must make valid predictions on the sample images.\nSecond, you need to make sure you are making valid predictions. \nI found that this was hard to achieve because my notebook threw exceptions when I submitted it. :(\nBut I am still trying. \nThe test folder contains 4000 per the description of the competition.\nI am still learning about code competition.\nI hope this helps.", "votes": null }, { "id": "2219459", "postDate": "04/12/2023 15:30:26", "content": "

Thank you for your answer!
\nI mean that I can't find out how to get prediction on test dataset technically.
\nWhat should I submit? Notebook with test prediction function in the last cellar or what?
\nIt's my first notebook-only competition that's why it's hard to understand such basic things.

", "rawMarkdown": "Thank you for your answer!\nI mean that I can't find out how to get prediction on test dataset technically.\nWhat should I submit? Notebook with test prediction function in the last cellar or what?\nIt's my first notebook-only competition that's why it's hard to understand such basic things.", "votes": null }, { "id": "2219660", "postDate": "04/12/2023 18:51:10", "content": "

You need to submit a notebook that at least works for the sample images and generates a CSV file that meets the required format. Check out this notebook https://www.kaggle.com/code/nbroad/donut-infer-lb-0-44-benetech
\nThis was my first notebook competition too. :)
\nI used to fork, submit, and didn't care about it.

", "rawMarkdown": "You need to submit a notebook that at least works for the sample images and generates a CSV file that meets the required format. Check out this notebook https://www.kaggle.com/code/nbroad/donut-infer-lb-0-44-benetech\nThis was my first notebook competition too. :)\nI used to fork, submit, and didn't care about it.", "votes": null }, { "id": "2219677", "postDate": "04/12/2023 19:16:37", "content": "

Here is the code requirement from the competition description.
\nhttps://www.kaggle.com/competitions/benetech-making-graphs-accessible/overview/code-requirements

", "rawMarkdown": "Here is the code requirement from the competition description.\nhttps://www.kaggle.com/competitions/benetech-making-graphs-accessible/overview/code-requirements", "votes": null }, { "id": "2220929", "postDate": "04/13/2023 19:29:56", "content": "

It makes a lot of sense for me now, thank you!

", "rawMarkdown": "It makes a lot of sense for me now, thank you!", "votes": null } ], "comments": [ { "id": 2218495, "author_name": "zhugds", "author_url": "", "post_date": "04/11/2023 18:39:01", "content": "

When you submit your notebook for scoring, it will make predictions on the real test images, which are different from the ones you see when you edit the notebook.
\nSo first, your notebook must make valid predictions on the sample images.
\nSecond, you need to make sure you are making valid predictions.
\nI found that this was hard to achieve because my notebook threw exceptions when I submitted it. :(
\nBut I am still trying.
\nThe test folder contains 4000 per the description of the competition.
\nI am still learning about code competition.
\nI hope this helps.

", "votes": null, "replies": [ { "id": 2219459, "author_name": "renatziganshin", "author_url": "", "post_date": "04/12/2023 15:30:26", "content": "

Thank you for your answer!
\nI mean that I can't find out how to get prediction on test dataset technically.
\nWhat should I submit? Notebook with test prediction function in the last cellar or what?
\nIt's my first notebook-only competition that's why it's hard to understand such basic things.

", "votes": null, "replies": [ { "id": 2219660, "author_name": "zhugds", "author_url": "", "post_date": "04/12/2023 18:51:10", "content": "

You need to submit a notebook that at least works for the sample images and generates a CSV file that meets the required format. Check out this notebook https://www.kaggle.com/code/nbroad/donut-infer-lb-0-44-benetech
\nThis was my first notebook competition too. :)
\nI used to fork, submit, and didn't care about it.

", "votes": null, "replies": [ { "id": 2220929, "author_name": "renatziganshin", "author_url": "", "post_date": "04/13/2023 19:29:56", "content": "

It makes a lot of sense for me now, thank you!

", "votes": null, "replies": [] } ] }, { "id": 2219677, "author_name": "zhugds", "author_url": "", "post_date": "04/12/2023 19:16:37", "content": "

Here is the code requirement from the competition description.
\nhttps://www.kaggle.com/competitions/benetech-making-graphs-accessible/overview/code-requirements

", "votes": null, "replies": [] } ] } ] } ], "raw_markdown_by_id": { "2215498": "How do I predict results on the test dataset in my notebook?\nMaybe its a bit silly question but I can not undetstand how to upload test images to my notebook for predictions.", "2218495": "When you submit your notebook for scoring, it will make predictions on the real test images, which are different from the ones you see when you edit the notebook. \nSo first, your notebook must make valid predictions on the sample images.\nSecond, you need to make sure you are making valid predictions. \nI found that this was hard to achieve because my notebook threw exceptions when I submitted it. :(\nBut I am still trying. \nThe test folder contains 4000 per the description of the competition.\nI am still learning about code competition.\nI hope this helps.", "2219459": "Thank you for your answer!\nI mean that I can't find out how to get prediction on test dataset technically.\nWhat should I submit? Notebook with test prediction function in the last cellar or what?\nIt's my first notebook-only competition that's why it's hard to understand such basic things.", "2219660": "You need to submit a notebook that at least works for the sample images and generates a CSV file that meets the required format. Check out this notebook https://www.kaggle.com/code/nbroad/donut-infer-lb-0-44-benetech\nThis was my first notebook competition too. :)\nI used to fork, submit, and didn't care about it.", "2219677": "Here is the code requirement from the competition description.\nhttps://www.kaggle.com/competitions/benetech-making-graphs-accessible/overview/code-requirements", "2220929": "It makes a lot of sense for me now, thank you!" }, "source": "meta" }