Are we going to be able to lets say retrain the model(s) once a week on the prior weeks data, or once a month on the prior weeks data or are we just to train the model up to the point in time it goes into qualification and after that, the model cannot be changed.
", "messages": [ { "id": "1569616", "postDate": "11/03/2021 15:31:58", "content": "Are we going to be able to lets say retrain the model(s) once a week on the prior weeks data, or once a month on the prior weeks data or are we just to train the model up to the point in time it goes into qualification and after that, the model cannot be changed.
", "rawMarkdown": "Are we going to be able to lets say retrain the model(s) once a week on the prior weeks data, or once a month on the prior weeks data or are we just to train the model up to the point in time it goes into qualification and after that, the model cannot be changed.", "votes": null }, { "id": "1569691", "postDate": "11/03/2021 16:53:21", "content": "You can retrain as much as you want with any dataset, given that the dataset is public and available to all
", "rawMarkdown": "You can retrain as much as you want with any dataset, given that the dataset is public and available to all", "votes": null }, { "id": "1569720", "postDate": "11/03/2021 17:16:22", "content": "So retraining during the leaderboard 3 month period is at our discression, Or is the model frozen at the end of the acceptance deadline?
", "rawMarkdown": "So retraining during the leaderboard 3 month period is at our discression, Or is the model frozen at the end of the acceptance deadline?", "votes": null }, { "id": "1569756", "postDate": "11/03/2021 17:29:08", "content": "You have to hand in a kernel, so the kernel is frozen. The question is if you will be able to re-train with new data in the kernel during reruns.
", "rawMarkdown": "You have to hand in a kernel, so the kernel is frozen. The question is if you will be able to re-train with new data in the kernel during reruns.", "votes": null }, { "id": "1569777", "postDate": "11/03/2021 17:54:43", "content": "Yeah sorry got the question wrong
", "rawMarkdown": "Yeah sorry got the question wrong", "votes": null }, { "id": "1570063", "postDate": "11/03/2021 23:52:01", "content": "No more new training data will be given by the host. You can use the given data and train models at anytime until the 3-month training phrase ends.
", "rawMarkdown": "No more new training data will be given by the host. You can use the given data and train models at anytime until the 3-month training phrase ends.", "votes": null }, { "id": "1570193", "postDate": "11/04/2021 01:18:55", "content": "Isn't \"supplemental_train.csv\" serving as additional training data after the training phase ends?
\n\n", "rawMarkdown": "Isn't \"supplemental_train.csv\" serving as additional training data after the training phase ends?\n> After the submission period is over this file's data will be replaced with cryptoasset prices from the submission period. The current copy, which is just filled approximately the right amount of data from train.csv is provided as a placeholder.", "votes": null }, { "id": "1570208", "postDate": "11/04/2021 01:33:56", "content": "After the submission period is over this file's data will be replaced with cryptoasset prices from the submission period. The current copy, which is just filled approximately the right amount of data from train.csv is provided as a placeholder.
\n
For supplemental_train.csv, I am still confused about the description. I don't know what the use of this file is. What's the significance of the file. That's why I made the post https://www.kaggle.com/c/g-research-crypto-forecasting/discussion/284946 .
\nWhether \"'supplemental_train.csv' serves as additional training data after the training phase ends\" by updating the data during the forecasting phrase needs to be clarified by the host.
I'm not as concerned about the training phase as much as the operational phase. Will there be any retraining allowed during operations?
", "rawMarkdown": "I'm not as concerned about the training phase as much as the operational phase. Will there be any retraining allowed during operations?", "votes": null }, { "id": "1572326", "postDate": "11/05/2021 15:42:41", "content": "@yus002 that was directed to you
", "rawMarkdown": "yus002 that was directed to you", "votes": null }, { "id": "1572375", "postDate": "11/05/2021 16:09:19", "content": "I have the same doubt, but right now it seems that the host will leave the question for us to figure out.
", "rawMarkdown": "I have the same doubt, but right now it seems that the host will leave the question for us to figure out.", "votes": null } ], "comments": [ { "id": 1569691, "author_name": "zarif98sjs", "author_url": "", "post_date": "11/03/2021 16:53:21", "content": "You can retrain as much as you want with any dataset, given that the dataset is public and available to all
", "votes": null, "replies": [ { "id": 1569756, "author_name": "philippsinger", "author_url": "", "post_date": "11/03/2021 17:29:08", "content": "You have to hand in a kernel, so the kernel is frozen. The question is if you will be able to re-train with new data in the kernel during reruns.
", "votes": null, "replies": [] }, { "id": 1569777, "author_name": "zarif98sjs", "author_url": "", "post_date": "11/03/2021 17:54:43", "content": "Yeah sorry got the question wrong
", "votes": null, "replies": [] } ] }, { "id": 1569720, "author_name": "rlucc001", "author_url": "", "post_date": "11/03/2021 17:16:22", "content": "So retraining during the leaderboard 3 month period is at our discression, Or is the model frozen at the end of the acceptance deadline?
", "votes": null, "replies": [] }, { "id": 1570063, "author_name": "yus002", "author_url": "", "post_date": "11/03/2021 23:52:01", "content": "No more new training data will be given by the host. You can use the given data and train models at anytime until the 3-month training phrase ends.
", "votes": null, "replies": [ { "id": 1570193, "author_name": "fritzcremer", "author_url": "", "post_date": "11/04/2021 01:18:55", "content": "Isn't \"supplemental_train.csv\" serving as additional training data after the training phase ends?
\n\n", "votes": null, "replies": [] }, { "id": 1570208, "author_name": "yus002", "author_url": "", "post_date": "11/04/2021 01:33:56", "content": "After the submission period is over this file's data will be replaced with cryptoasset prices from the submission period. The current copy, which is just filled approximately the right amount of data from train.csv is provided as a placeholder.
\n
For supplemental_train.csv, I am still confused about the description. I don't know what the use of this file is. What's the significance of the file. That's why I made the post https://www.kaggle.com/c/g-research-crypto-forecasting/discussion/284946 .
\nWhether \"'supplemental_train.csv' serves as additional training data after the training phase ends\" by updating the data during the forecasting phrase needs to be clarified by the host.
I'm not as concerned about the training phase as much as the operational phase. Will there be any retraining allowed during operations?
", "votes": null, "replies": [ { "id": 1572375, "author_name": "yus002", "author_url": "", "post_date": "11/05/2021 16:09:19", "content": "I have the same doubt, but right now it seems that the host will leave the question for us to figure out.
", "votes": null, "replies": [] } ] }, { "id": 1572326, "author_name": "rlucc001", "author_url": "", "post_date": "11/05/2021 15:42:41", "content": "@yus002 that was directed to you
", "votes": null, "replies": [] } ], "raw_markdown_by_id": { "1569616": "Are we going to be able to lets say retrain the model(s) once a week on the prior weeks data, or once a month on the prior weeks data or are we just to train the model up to the point in time it goes into qualification and after that, the model cannot be changed.", "1569691": "You can retrain as much as you want with any dataset, given that the dataset is public and available to all", "1569720": "So retraining during the leaderboard 3 month period is at our discression, Or is the model frozen at the end of the acceptance deadline?", "1569756": "You have to hand in a kernel, so the kernel is frozen. The question is if you will be able to re-train with new data in the kernel during reruns.", "1569777": "Yeah sorry got the question wrong", "1570063": "No more new training data will be given by the host. You can use the given data and train models at anytime until the 3-month training phrase ends.", "1570193": "Isn't \"supplemental_train.csv\" serving as additional training data after the training phase ends?\n> After the submission period is over this file's data will be replaced with cryptoasset prices from the submission period. The current copy, which is just filled approximately the right amount of data from train.csv is provided as a placeholder.", "1570208": "For supplemental_train.csv, I am still confused about the description. I don't know what the use of this file is. What's the significance of the file. That's why I made the post https://www.kaggle.com/c/g-research-crypto-forecasting/discussion/284946 .\nWhether \"'supplemental_train.csv' serves as additional training data after the training phase ends\" by updating the data during the forecasting phrase needs to be clarified by the host.", "1572324": "I'm not as concerned about the training phase as much as the operational phase. Will there be any retraining allowed during operations?", "1572326": "yus002 that was directed to you", "1572375": "I have the same doubt, but right now it seems that the host will leave the question for us to figure out." }, "source": "meta" }