mmA is around 0.40 on the training set, but the score on the test set after submission is only 0.001. The code runs fine on the training set. What could be causing this?
", "messages": [ { "id": "2244109", "postDate": "05/03/2023 12:54:57", "content": "mmA is around 0.40 on the training set, but the score on the test set after submission is only 0.001. The code runs fine on the training set. What could be causing this?
", "rawMarkdown": "mmA is around 0.40 on the training set, but the score on the test set after submission is only 0.001. The code runs fine on the training set. What could be causing this?", "votes": null }, { "id": "2244116", "postDate": "05/03/2023 13:00:44", "content": "I modified the path of the test data according to the sample_submission, and the code runs successfully without taking more than 8 hours for evaluation.
", "rawMarkdown": "I modified the path of the test data according to the sample_submission, and the code runs successfully without taking more than 8 hours for evaluation.", "votes": null }, { "id": "2244211", "postDate": "05/03/2023 14:08:55", "content": "That almost certainly looks like an issue with your script, because the results are essentially the same as random numbers. You might not be overwriting the csv rows correctly: double-check your code against our submission example, maybe?
", "rawMarkdown": "That almost certainly looks like an issue with your script, because the results are essentially the same as random numbers. You might not be overwriting the csv rows correctly: double-check your code against our submission example, maybe?", "votes": null }, { "id": "2244305", "postDate": "05/03/2023 15:05:02", "content": "Thank you! I'll check my code again. By the way, is the size of the test set much different from that of the training set, or are they roughly the same size（number of images）?
", "rawMarkdown": "Thank you! I'll check my code again. By the way, is the size of the test set much different from that of the training set, or are they roughly the same size（number of images）?", "votes": null }, { "id": "2244432", "postDate": "05/03/2023 16:16:11", "content": "See this page: \"Expect to find roughly 1,100 images in the hidden test set. The number of images in a scene may vary from <10 to ~250.\"
", "rawMarkdown": "See [this page](https://www.kaggle.com/competitions/image-matching-challenge-2023/data): \"Expect to find roughly 1,100 images in the hidden test set. The number of images in a scene may vary from <10 to ~250.\"", "votes": null }, { "id": "2245633", "postDate": "05/04/2023 13:48:25", "content": "I don't think this is a personal fault. I also met this problem.
\nAnd I found that once I add a empty function into official code——https://www.kaggle.com/code/eduardtrulls/imc-2023-submission-example, the score fell down from 0.04 to 0.001
here is the code I used : https://www.kaggle.com/code/khuang724/imc-2023-submission-example-89d173,
\nafter annotating the function 'process()', the score increase to 0.04 from 0.001
https://www.kaggle.com/code/khuang724/add-a-empty-function-and-get-score-0-001
", "rawMarkdown": "https://www.kaggle.com/code/khuang724/add-a-empty-function-and-get-score-0-001", "votes": null }, { "id": "2245827", "postDate": "05/04/2023 16:04:37", "content": "I'm not sure what you're trying to do. Note that the code is wrapped into a try-except block to be more robust against unplanned errors. As far as I can tell your loop calls the process function with undefined parameters (e.g. feature_file), so it will just fail?
I modified the path of the test data according to the sample_submission, and the code runs successfully without taking more than 8 hours for evaluation.
", "votes": null, "replies": [ { "id": 2244211, "author_name": "eduardtrulls", "author_url": "", "post_date": "05/03/2023 14:08:55", "content": "That almost certainly looks like an issue with your script, because the results are essentially the same as random numbers. You might not be overwriting the csv rows correctly: double-check your code against our submission example, maybe?
", "votes": null, "replies": [ { "id": 2244305, "author_name": "arphei", "author_url": "", "post_date": "05/03/2023 15:05:02", "content": "Thank you! I'll check my code again. By the way, is the size of the test set much different from that of the training set, or are they roughly the same size（number of images）?
", "votes": null, "replies": [ { "id": 2244432, "author_name": "eduardtrulls", "author_url": "", "post_date": "05/03/2023 16:16:11", "content": "See this page: \"Expect to find roughly 1,100 images in the hidden test set. The number of images in a scene may vary from <10 to ~250.\"
", "votes": null, "replies": [] } ] }, { "id": 2245633, "author_name": "khuang724", "author_url": "", "post_date": "05/04/2023 13:48:25", "content": "I don't think this is a personal fault. I also met this problem.
\nAnd I found that once I add a empty function into official code——https://www.kaggle.com/code/eduardtrulls/imc-2023-submission-example, the score fell down from 0.04 to 0.001
here is the code I used : https://www.kaggle.com/code/khuang724/imc-2023-submission-example-89d173,
\nafter annotating the function 'process()', the score increase to 0.04 from 0.001
https://www.kaggle.com/code/khuang724/add-a-empty-function-and-get-score-0-001
", "votes": null, "replies": [ { "id": 2245827, "author_name": "eduardtrulls", "author_url": "", "post_date": "05/04/2023 16:04:37", "content": "I'm not sure what you're trying to do. Note that the code is wrapped into a try-except block to be more robust against unplanned errors. As far as I can tell your loop calls the process function with undefined parameters (e.g. feature_file), so it will just fail?