{ "id": 216871, "title": "Unable to Submit Successfully", "url": "/competitions/cassava-leaf-disease-classification/discussion/216871", "author_name": "", "post_date": "2021-02-04T10:57:46.300697700Z", "votes": -1, "comment_count": 5, "views": 0, "content": "

Can anyone help me with how to submit successfully to the competition. This is my first ever time using kaggle and it's proving difficult for me. I trained on a notebook using TPU and saved my model, which I then called in another notebook using GPU for inference. I have been able to generate the submission.csv file but after submitting I keep on getting the error submission scoring error. What am I doing wrong. Here are image links showing the error and code used to create the submission file

", "messages": [ { "id": "1185759", "postDate": "02/04/2021 10:57:46", "content": "

Can anyone help me with how to submit successfully to the competition. This is my first ever time using kaggle and it's proving difficult for me. I trained on a notebook using TPU and saved my model, which I then called in another notebook using GPU for inference. I have been able to generate the submission.csv file but after submitting I keep on getting the error submission scoring error. What am I doing wrong. Here are image links showing the error and code used to create the submission file

", "rawMarkdown": "Can anyone help me with how to submit successfully to the competition. This is my first ever time using kaggle and it's proving difficult for me. I trained on a notebook using TPU and saved my model, which I then called in another notebook using GPU for inference. I have been able to generate the **submission.csv** file but after submitting I keep on getting the error **submission scoring error.** What am I doing wrong. Here are image links showing the [error](https://drive.google.com/file/d/15VGq3YpCNKKGUsKGlrDUb9HAXan3LahR/view?usp=sharing) and code used to create the [submission file](https://drive.google.com/file/d/1_BLIQ-WEDBIcEUGVZkhbFQeiwvHTYx7L/view?usp=sharing)", "votes": null }, { "id": "1186071", "postDate": "02/04/2021 15:15:44", "content": "

There are a number of posts in this competition and in several other vision competitions in the past year where this method of generating submissions is discussed.

\n

A significant tool that data scientists must learn is the ability to perform searches. For fixing Python code Google is for sure a very good tool. For fixing kaggle related issues the kaggle search (cute little magnifying glass) is great place to start by searching posts in this competition and than searching in the full site.

\n

You had one down vote when I read your question - guessing it was the result of your apparent failure to search posts :)

\n

For any real help with this type of error you need to make your kernels public so we can see your complete code. The little snip you provided is almost of no value.

", "rawMarkdown": "There are a number of posts in this competition and in several other vision competitions in the past year where this method of generating submissions is discussed. \n\nA significant tool that data scientists must learn is the ability to perform searches. For fixing Python code Google is for sure a very good tool. For fixing kaggle related issues the kaggle search (cute little magnifying glass) is great place to start by searching posts in this competition and than searching in the full site.\n\nYou had one down vote when I read your question - guessing it was the result of your apparent failure to search posts :)\n\nFor any real help with this type of error you need to make your kernels public so we can see your complete code. The little snip you provided is almost of no value.", "votes": null }, { "id": "1186148", "postDate": "02/04/2021 15:56:01", "content": "

Much grateful for your candid feedback. I will do well to up my skills

", "rawMarkdown": "Much grateful for your candid feedback. I will do well to up my skills", "votes": null }, { "id": "1186285", "postDate": "02/04/2021 18:14:36", "content": "

I have created 2 noteboooks for this very specific purpose. You can find details here

", "rawMarkdown": "I have created 2 noteboooks for this very specific purpose. You can find details [here](https://www.kaggle.com/c/cassava-leaf-disease-classification/discussion/216461)", "votes": null }, { "id": "1187223", "postDate": "02/05/2021 09:54:45", "content": "

This worked. Thank you very much.

\n

This also helped me figure out the error. While I used the .tfrecords for training, I wasn't using that for inference instead used the image in the test_images folder. Even though the inference worked the submission might have \"realised\" that difference and hence spewed out errors.

", "rawMarkdown": "This worked. Thank you very much. \n\nThis also helped me figure out the error. While I used the .tfrecords for training, I wasn't using that for inference instead used the image in the test_images folder. Even though the inference worked the submission might have \"realised\" that difference and hence spewed out errors.", "votes": null }, { "id": "1187272", "postDate": "02/05/2021 10:28:48", "content": "

Great, happy to help. The notebooks just need to be tweaked more or less to make it work for almost any similar problem.

", "rawMarkdown": "Great, happy to help. The notebooks just need to be tweaked more or less to make it work for almost any similar problem.", "votes": null } ], "comments": [ { "id": 1186071, "author_name": "pcjimmmy", "author_url": "", "post_date": "02/04/2021 15:15:44", "content": "

There are a number of posts in this competition and in several other vision competitions in the past year where this method of generating submissions is discussed.

\n

A significant tool that data scientists must learn is the ability to perform searches. For fixing Python code Google is for sure a very good tool. For fixing kaggle related issues the kaggle search (cute little magnifying glass) is great place to start by searching posts in this competition and than searching in the full site.

\n

You had one down vote when I read your question - guessing it was the result of your apparent failure to search posts :)

\n

For any real help with this type of error you need to make your kernels public so we can see your complete code. The little snip you provided is almost of no value.

", "votes": null, "replies": [ { "id": 1186148, "author_name": "atiaisaac", "author_url": "", "post_date": "02/04/2021 15:56:01", "content": "

Much grateful for your candid feedback. I will do well to up my skills

", "votes": null, "replies": [] } ] }, { "id": 1186285, "author_name": "vickygoyal", "author_url": "", "post_date": "02/04/2021 18:14:36", "content": "

I have created 2 noteboooks for this very specific purpose. You can find details here

", "votes": null, "replies": [ { "id": 1187223, "author_name": "atiaisaac", "author_url": "", "post_date": "02/05/2021 09:54:45", "content": "

This worked. Thank you very much.

\n

This also helped me figure out the error. While I used the .tfrecords for training, I wasn't using that for inference instead used the image in the test_images folder. Even though the inference worked the submission might have \"realised\" that difference and hence spewed out errors.

", "votes": null, "replies": [] }, { "id": 1187272, "author_name": "vickygoyal", "author_url": "", "post_date": "02/05/2021 10:28:48", "content": "

Great, happy to help. The notebooks just need to be tweaked more or less to make it work for almost any similar problem.

", "votes": null, "replies": [] } ] } ], "raw_markdown_by_id": { "1185759": "Can anyone help me with how to submit successfully to the competition. This is my first ever time using kaggle and it's proving difficult for me. I trained on a notebook using TPU and saved my model, which I then called in another notebook using GPU for inference. I have been able to generate the **submission.csv** file but after submitting I keep on getting the error **submission scoring error.** What am I doing wrong. Here are image links showing the [error](https://drive.google.com/file/d/15VGq3YpCNKKGUsKGlrDUb9HAXan3LahR/view?usp=sharing) and code used to create the [submission file](https://drive.google.com/file/d/1_BLIQ-WEDBIcEUGVZkhbFQeiwvHTYx7L/view?usp=sharing)", "1186071": "There are a number of posts in this competition and in several other vision competitions in the past year where this method of generating submissions is discussed. \n\nA significant tool that data scientists must learn is the ability to perform searches. For fixing Python code Google is for sure a very good tool. For fixing kaggle related issues the kaggle search (cute little magnifying glass) is great place to start by searching posts in this competition and than searching in the full site.\n\nYou had one down vote when I read your question - guessing it was the result of your apparent failure to search posts :)\n\nFor any real help with this type of error you need to make your kernels public so we can see your complete code. The little snip you provided is almost of no value.", "1186148": "Much grateful for your candid feedback. I will do well to up my skills", "1186285": "I have created 2 noteboooks for this very specific purpose. You can find details [here](https://www.kaggle.com/c/cassava-leaf-disease-classification/discussion/216461)", "1187223": "This worked. Thank you very much. \n\nThis also helped me figure out the error. While I used the .tfrecords for training, I wasn't using that for inference instead used the image in the test_images folder. Even though the inference worked the submission might have \"realised\" that difference and hence spewed out errors.", "1187272": "Great, happy to help. The notebooks just need to be tweaked more or less to make it work for almost any similar problem." }, "source": "meta" }