![\"\"](\"https://www.googleapis.com/download/storage/v1/b/kaggle-forum-message-attachments/o/inbox%2F5387387%2Fd9cf6acf0f3e06f43df7c55861520e24%2Fkaggle%20submission%2020230206.png?generation=1675669077871602&alt=media\")
This submission completes successfully but receives a score of 0
\n
The confusion matrix derived from the training set
\n![\"\"](\"https://www.googleapis.com/download/storage/v1/b/kaggle-forum-message-attachments/o/inbox%2F5387387%2Fe642fa80671eb8905bc543d78ca16955%2Fkaggle%20confusion%20resnet18%2020230205.png?generation=1675669198063362&alt=media\")
suggests the F1 score should be reasonable even though the sample is unbalanced
\n![\"\"](\"https://www.googleapis.com/download/storage/v1/b/kaggle-forum-message-attachments/o/inbox%2F5387387%2F2d295ccb4d64ded7802feaa27a193995%2Fkaggle%20F1%2020230206.png?generation=1675669763172195&alt=media\")
Any advice anybody can offer on why this is happening and how to resolve would be much appreciated
\nMany thanks
\nJulian
", "messages": [ { "id": 2131875, "postDate": "2023-02-06T13:02:20.377Z", "content": "I think the problem was that when I created my augmented images, the image names were not being updated correctly so I ended up with a heap of duplicates. I now have a score - albeit a very low one which makes this an even more interesting challenge than I thought originally
", "rawMarkdown": "I think the problem was that when I created my augmented images, the image names were not being updated correctly so I ended up with a heap of duplicates. I now have a score - albeit a very low one which makes this an even more interesting challenge than I thought originally", "votes": 1 }, { "id": 2131610, "postDate": "2023-02-06T08:28:11.110Z", "content": "If all the values under the column 'cancer' in the submission.csv file are zero, the score becomes zero. You might have a small bug in your code where it generates the submission csv file?
", "rawMarkdown": "If all the values under the column 'cancer' in the submission.csv file are zero, the score becomes zero. You might have a small bug in your code where it generates the submission csv file?", "votes": 1, "replies": [ { "id": 2131866, "postDate": "2023-02-06T12:58:34.977Z", "content": "Many thanks for the insight
\nAll the best
\nJulian
This submission completes successfully but receives a score of 0
\n
The confusion matrix derived from the training set
\n
suggests the F1 score should be reasonable even though the sample is unbalanced
\n
Any advice anybody can offer on why this is happening and how to resolve would be much appreciated
\nMany thanks
\nJulian
", "rawMarkdown": "[This submission](https://www.kaggle.com/code/julianmacnamara/rsna-breast-cancer-detection-pttoa-v1-1) completes successfully but receives a score of 0\n\n\n\nThe confusion matrix derived from the training set\n\n\n\nsuggests the F1 score should be reasonable even though the sample is unbalanced\n\n\n\nAny advice anybody can offer on why this is happening and how to resolve would be much appreciated\n\nMany thanks\n\nJulian", "votes": 1 }, { "id": 2131622, "postDate": "2023-02-06T08:35:09.427Z", "content": "For me score zero happens when model output has nan’s, try to check your output for nan’s
", "rawMarkdown": "For me score zero happens when model output has nan’s, try to check your output for nan’s ", "replies": [ { "id": 2131864, "postDate": "2023-02-06T12:57:37.697Z", "content": "Many thanks for the suggestion
\nAll the best
\nJulian
I think the problem was that when I created my augmented images, the image names were not being updated correctly so I ended up with a heap of duplicates. I now have a score - albeit a very low one which makes this an even more interesting challenge than I thought originally
", "votes": 1, "replies": [] }, { "id": 2131610, "author_name": "Rasoul Mojtahedzadeh", "author_url": "", "post_date": "2023-02-06T08:28:11.110000", "content": "If all the values under the column 'cancer' in the submission.csv file are zero, the score becomes zero. You might have a small bug in your code where it generates the submission csv file?
", "votes": 1, "replies": [ { "id": 2131866, "author_name": "Julian Macnamara", "author_url": "", "post_date": "2023-02-06T12:58:34.977000", "content": "Many thanks for the insight
\nAll the best
\nJulian
For me score zero happens when model output has nan’s, try to check your output for nan’s
", "votes": 0, "replies": [ { "id": 2131864, "author_name": "Julian Macnamara", "author_url": "", "post_date": "2023-02-06T12:57:37.697000", "content": "Many thanks for the suggestion
\nAll the best
\nJulian