Can anybody please tell me how to make correct submission I have trained model and store it .
", "messages": [ { "id": "3045826", "postDate": "11/14/2024 20:50:19", "content": "Can anybody please tell me how to make correct submission I have trained model and store it .
", "rawMarkdown": "Can anybody please tell me how to make correct submission I have trained model and store it .", "votes": null }, { "id": "3045827", "postDate": "11/14/2024 20:50:59", "content": "I have tried submission other people forked notebook and try to understand it but i have not figured out my mistake.
", "rawMarkdown": "I have tried submission other people forked notebook and try to understand it but i have not figured out my mistake.", "votes": null }, { "id": "3046053", "postDate": "11/15/2024 04:56:14", "content": "@muhammadqasimshabbir
\nFork the notebook here and replace the submission constant with your model. Something like this will work-
preds = model.predict(yourtestdataset)\npredictions.select(pl.Series(, test[]).with_columns(pl.Series(, preds)).write_parquet()\nthanks i will give it a shot.
", "rawMarkdown": "thanks i will give it a shot.", "votes": null }, { "id": "3046481", "postDate": "11/15/2024 14:11:33", "content": "I have a question here could you please tell me your test_data how to is this we have to do the same preprocessing in the def predict(test: pl.DataFrame, lags: pl.DataFrame | None) -> pl.DataFrame | pd.DataFrame: inside this funtion i have to preprocess test data same as in training then I have to input this dataframe to model for predcition in above code . Am right or not ?
", "rawMarkdown": "I have a question here could you please tell me your test_data how to is this we have to do the same preprocessing in the ***def predict(test: pl.DataFrame, lags: pl.DataFrame | None) -> pl.DataFrame | pd.DataFrame:*** inside this funtion i have to preprocess test data same as in training then I have to input this dataframe to model for predcition in above code . Am right or not ?", "votes": null }, { "id": "3046986", "postDate": "11/16/2024 05:47:55", "content": "Yes of course, this is the same as any other pipeline - you create the preprocessing for the test set batch like the one used in the train set. This is then used for prediction and a new batch is served @muhammadqasimshabbir
", "rawMarkdown": "Yes of course, this is the same as any other pipeline - you create the preprocessing for the test set batch like the one used in the train set. This is then used for prediction and a new batch is served @muhammadqasimshabbir", "votes": null }, { "id": "3047404", "postDate": "11/16/2024 16:23:23", "content": "Thanks for your help @ravi20076
", "rawMarkdown": "Thanks for your help @ravi20076", "votes": null } ], "comments": [ { "id": 3045827, "author_name": "", "author_url": "", "post_date": "11/14/2024 20:50:59", "content": "I have tried submission other people forked notebook and try to understand it but i have not figured out my mistake.
", "votes": null, "replies": [ { "id": 3046053, "author_name": "ravi20076", "author_url": "", "post_date": "11/15/2024 04:56:14", "content": "@muhammadqasimshabbir
\nFork the notebook here and replace the submission constant with your model. Something like this will work-
preds = model.predict(yourtestdataset)\npredictions.select(pl.Series(, test[]).with_columns(pl.Series(, preds)).write_parquet()\nthanks i will give it a shot.
", "votes": null, "replies": [ { "id": 3046481, "author_name": "", "author_url": "", "post_date": "11/15/2024 14:11:33", "content": "I have a question here could you please tell me your test_data how to is this we have to do the same preprocessing in the def predict(test: pl.DataFrame, lags: pl.DataFrame | None) -> pl.DataFrame | pd.DataFrame: inside this funtion i have to preprocess test data same as in training then I have to input this dataframe to model for predcition in above code . Am right or not ?
", "votes": null, "replies": [ { "id": 3046986, "author_name": "ravi20076", "author_url": "", "post_date": "11/16/2024 05:47:55", "content": "Yes of course, this is the same as any other pipeline - you create the preprocessing for the test set batch like the one used in the train set. This is then used for prediction and a new batch is served @muhammadqasimshabbir
", "votes": null, "replies": [ { "id": 3047404, "author_name": "", "author_url": "", "post_date": "11/16/2024 16:23:23", "content": "Thanks for your help @ravi20076
", "votes": null, "replies": [] } ] } ] } ] } ] } ] } ], "raw_markdown_by_id": { "3045826": "Can anybody please tell me how to make correct submission I have trained model and store it .", "3045827": "I have tried submission other people forked notebook and try to understand it but i have not figured out my mistake.", "3046053": "muhammadqasimshabbir \nFork the notebook [here](https://www.kaggle.com/code/ryanholbrook/jane-street-rmf-demo-submission) and replace the submission constant with your model. Something like this will work-\n\n```python\npreds = model.predict(yourtestdataset)\npredictions.select(pl.Series(\"row_id\", test[\"row_id\"]).with_columns(pl.Series(\"responder_6\", preds)).write_parquet(\"submission.parquet\")\n```", "3046479": "thanks i will give it a shot.", "3046481": "I have a question here could you please tell me your test_data how to is this we have to do the same preprocessing in the ***def predict(test: pl.DataFrame, lags: pl.DataFrame | None) -> pl.DataFrame | pd.DataFrame:*** inside this funtion i have to preprocess test data same as in training then I have to input this dataframe to model for predcition in above code . Am right or not ?", "3046986": "Yes of course, this is the same as any other pipeline - you create the preprocessing for the test set batch like the one used in the train set. This is then used for prediction and a new batch is served @muhammadqasimshabbir", "3047404": "Thanks for your help @ravi20076" }, "source": "meta" }