{ "id": 574085, "title": "submission.csv file not found", "url": "/competitions/stanford-rna-3d-folding/discussion/574085", "author_name": "", "post_date": "2025-04-19T20:29:47.095318300Z", "votes": 2, "comment_count": 2, "views": 0, "content": "

hey, i am stuck at submissions. Whenever i try to submit NB for scoring it gives error say submission.csv does not existed, although the file is created during inference. Can u guys pls tell me what causing that error.
\n\"\"

\n

\"\"

\n

here is the notebook!
\nhttps://www.kaggle.com/code/shreyasraghav/modified-dataset-with-parquet-files

", "messages": [ { "id": "3182785", "postDate": "04/19/2025 20:29:47", "content": "

hey, i am stuck at submissions. Whenever i try to submit NB for scoring it gives error say submission.csv does not existed, although the file is created during inference. Can u guys pls tell me what causing that error.
\n\"\"

\n

\"\"

\n

here is the notebook!
\nhttps://www.kaggle.com/code/shreyasraghav/modified-dataset-with-parquet-files

", "rawMarkdown": "hey, i am stuck at submissions. Whenever i try to submit NB for scoring it gives error say submission.csv does not existed, although the file is created during inference. Can u guys pls tell me what causing that error.\n![](https://www.googleapis.com/download/storage/v1/b/kaggle-forum-message-attachments/o/inbox%2F20808938%2F0ce445168029b3efb50dbdda6fbd279f%2FScreenshot%202025-04-20%20015428.png?generation=1745094542764177&alt=media)\n\n![](https://www.googleapis.com/download/storage/v1/b/kaggle-forum-message-attachments/o/inbox%2F20808938%2Fef14021a2af0ca72c1a5fce3880fd4ce%2FScreenshot%202025-04-20%20015444.png?generation=1745094568874694&alt=media)\n\nhere is the notebook!\nhttps://www.kaggle.com/code/shreyasraghav/modified-dataset-with-parquet-files", "votes": null }, { "id": "3182792", "postDate": "04/19/2025 20:55:57", "content": "

I noticed you're using %%writefile model.py to save your model definition. While this can work during public test runs or short submissions, it often fails during private or hidden test evaluations. This is because the file might not persist correctly or may not be found by the environment at runtime.

\n

To avoid this issue, it's better to define the model directly in the notebook rather than writing it to a .py file. This ensures the model is always accessible and avoids potential import errors.

\n

Alternatively, if you really want to keep your model code modular:

\n", "rawMarkdown": "I noticed you're using %%writefile model.py to save your model definition. While this can work during public test runs or short submissions, it often fails during private or hidden test evaluations. This is because the file might not persist correctly or may not be found by the environment at runtime.\n\nTo avoid this issue, it's better to define the model directly in the notebook rather than writing it to a .py file. This ensures the model is always accessible and avoids potential import errors.\n\nAlternatively, if you really want to keep your model code modular:\n- Save the full model code in a model.py file beforehand on your local machine.\n- Upload it to your Kaggle notebook using the \"Add file\" option.\n- Then, in your notebook, use the following code to load it reliably:\n \n import sys\n sys.path.append('/kaggle/working') # or the directory where you uploaded the model.py\n from model import MyModel # or whatever your class/function is called", "votes": null }, { "id": "3182795", "postDate": "04/19/2025 21:00:10", "content": "

thanks sir, I hope this will work out.

", "rawMarkdown": "thanks sir, I hope this will work out.", "votes": null } ], "comments": [ { "id": 3182792, "author_name": "arunodhayan", "author_url": "", "post_date": "04/19/2025 20:55:57", "content": "

I noticed you're using %%writefile model.py to save your model definition. While this can work during public test runs or short submissions, it often fails during private or hidden test evaluations. This is because the file might not persist correctly or may not be found by the environment at runtime.

\n

To avoid this issue, it's better to define the model directly in the notebook rather than writing it to a .py file. This ensures the model is always accessible and avoids potential import errors.

\n

Alternatively, if you really want to keep your model code modular:

\n", "votes": null, "replies": [ { "id": 3182795, "author_name": "shreyasraghav", "author_url": "", "post_date": "04/19/2025 21:00:10", "content": "

thanks sir, I hope this will work out.

", "votes": null, "replies": [] } ] } ], "raw_markdown_by_id": { "3182785": "hey, i am stuck at submissions. Whenever i try to submit NB for scoring it gives error say submission.csv does not existed, although the file is created during inference. Can u guys pls tell me what causing that error.\n![](https://www.googleapis.com/download/storage/v1/b/kaggle-forum-message-attachments/o/inbox%2F20808938%2F0ce445168029b3efb50dbdda6fbd279f%2FScreenshot%202025-04-20%20015428.png?generation=1745094542764177&alt=media)\n\n![](https://www.googleapis.com/download/storage/v1/b/kaggle-forum-message-attachments/o/inbox%2F20808938%2Fef14021a2af0ca72c1a5fce3880fd4ce%2FScreenshot%202025-04-20%20015444.png?generation=1745094568874694&alt=media)\n\nhere is the notebook!\nhttps://www.kaggle.com/code/shreyasraghav/modified-dataset-with-parquet-files", "3182792": "I noticed you're using %%writefile model.py to save your model definition. While this can work during public test runs or short submissions, it often fails during private or hidden test evaluations. This is because the file might not persist correctly or may not be found by the environment at runtime.\n\nTo avoid this issue, it's better to define the model directly in the notebook rather than writing it to a .py file. This ensures the model is always accessible and avoids potential import errors.\n\nAlternatively, if you really want to keep your model code modular:\n- Save the full model code in a model.py file beforehand on your local machine.\n- Upload it to your Kaggle notebook using the \"Add file\" option.\n- Then, in your notebook, use the following code to load it reliably:\n \n import sys\n sys.path.append('/kaggle/working') # or the directory where you uploaded the model.py\n from model import MyModel # or whatever your class/function is called", "3182795": "thanks sir, I hope this will work out." }, "source": "meta" }