![\"\"](\"https://www.googleapis.com/download/storage/v1/b/kaggle-forum-message-attachments/o/inbox%2F4001300%2F439e88532c3b4d0fd8ec963f30df67a0%2FClipboard_05-31-2025_01.jpg?generation=1748622042046780&alt=media\")
Unlike typical competitions, a sample_submission.csv is not provided for this one.
\nInstead, you will make submissions in a way similar to what might be described on \"this notebook\" (referring to the competition's specific submission instructions page).
https://www.kaggle.com/code/sohier/cmi-2025-demo-submission
\nSince this can be a bit confusing, I will briefly explain it for beginners.
\nFirst, as you can see from the train and test data, one target (gesture) is determined for each sequence ID.
\n
Then, for submission, in the def predict function on the above notebook, data for each sequence will be sequentially provided as Polars DataFrames under the variable names sequence and demographics.
\n![\"\"](\"https://www.googleapis.com/download/storage/v1/b/kaggle-forum-message-attachments/o/inbox%2F4001300%2F74b27d36503c18daa238890efd4656a0%2FClipboard_05-31-2025_02.jpg?generation=1748622178296687&alt=media\")
You need to predict the gesture for that sequence and return the predicted gesture as text.
\n\"Sequence by sequence\" means that once you return a predicted value for one sequence, the next sequence will be provided, and you will then make a prediction for that one, repeating this process.
\nJudging from the metric URL, it seems that something like a submission.csv is automatically generated internally, but we no longer need to create it ourselves.
\nEnjoy kaggling!
", "messages": [ { "id": "3213892", "postDate": "05/30/2025 16:24:40", "content": "Unlike typical competitions, a sample_submission.csv is not provided for this one.
\nInstead, you will make submissions in a way similar to what might be described on \"this notebook\" (referring to the competition's specific submission instructions page).
https://www.kaggle.com/code/sohier/cmi-2025-demo-submission
\nSince this can be a bit confusing, I will briefly explain it for beginners.
\nFirst, as you can see from the train and test data, one target (gesture) is determined for each sequence ID.
\n
Then, for submission, in the def predict function on the above notebook, data for each sequence will be sequentially provided as Polars DataFrames under the variable names sequence and demographics.
\n
You need to predict the gesture for that sequence and return the predicted gesture as text.
\n\"Sequence by sequence\" means that once you return a predicted value for one sequence, the next sequence will be provided, and you will then make a prediction for that one, repeating this process.
\nJudging from the metric URL, it seems that something like a submission.csv is automatically generated internally, but we no longer need to create it ourselves.
\nEnjoy kaggling!
", "rawMarkdown": "Unlike typical competitions, a sample_submission.csv is not provided for this one.\nInstead, you will make submissions in a way similar to what might be described on \"this notebook\" (referring to the competition's specific submission instructions page). \n\nhttps://www.kaggle.com/code/sohier/cmi-2025-demo-submission\n\n\nSince this can be a bit confusing, I will briefly explain it for beginners.\n\nFirst, as you can see from the train and test data, one target (gesture) is determined for each sequence ID.\n\n\n\nThen, for submission, in the def predict function on the above notebook, data for each sequence will be sequentially provided as Polars DataFrames under the variable names sequence and demographics.\n\n\n\nYou need to predict the gesture for that sequence and return the predicted gesture as text.\n\n \"Sequence by sequence\" means that once you return a predicted value for one sequence, the next sequence will be provided, and you will then make a prediction for that one, repeating this process.\n\nJudging from the metric URL, it seems that something like a submission.csv is automatically generated internally, but we no longer need to create it ourselves.\n\nEnjoy kaggling!", "votes": null }, { "id": "3213914", "postDate": "05/30/2025 16:57:22", "content": "Submission File\n\nYou must submit to this competition using the provided evaluation API, which ensures that models perform inference on a single sequence at a time. For each sequence_id the test , you must predict the corresponding gesture.\n![\"\"](\"https://www.googleapis.com/download/storage/v1/b/kaggle-forum-message-attachments/o/inbox%2F23276411%2F7a2dd884d38b11cb3fcd5ae320a3f015%2FScreenshot%20from%202025-05-30%2022-24-39.png?generation=1748624129319786&alt=media\")
@sangrampatil5150 Thank you so much!!
", "rawMarkdown": "sangrampatil5150 Thank you so much!!", "votes": null }, { "id": "3217812", "postDate": "06/05/2025 12:24:30", "content": "Thanks for such instructions! I still have a question: for the Polars DataFrame that will be passed into predict function, what will it look like? Will it be the same as given test.csv data?🤔
Submission File\n\nYou must submit to this competition using the provided evaluation API, which ensures that models perform inference on a single sequence at a time. For each sequence_id the test , you must predict the corresponding gesture.\n
@sangrampatil5150 Thank you so much!!
", "votes": null, "replies": [] } ] }, { "id": 3217812, "author_name": "zequanbear", "author_url": "", "post_date": "06/05/2025 12:24:30", "content": "Thanks for such instructions! I still have a question: for the Polars DataFrame that will be passed into predict function, what will it look like? Will it be the same as given test.csv data?🤔