{ "id": 311937, "title": "Important: Prize Eligibility", "url": "/competitions/ultra-mnist/discussion/311937", "author_name": "", "post_date": "2022-03-09T15:02:19.215676900Z", "votes": 25, "comment_count": 11, "views": 0, "content": "

In order to be eligible for the prizes, you must:

\n", "messages": [ { "id": "1716988", "postDate": "03/09/2022 15:02:19", "content": "

In order to be eligible for the prizes, you must:

\n", "rawMarkdown": "In order to be eligible for the prizes, you must:\n\n- Register for NVIDIA GTC using this link: https://www.nvidia.com/gtc/?ncid=ref-crea-864532\n\n- Attend GTC sessions. Note: Prizes will be awarded only to those who register for GTC using the link above and attend some sessions.\n\n- Winners must open-source their solution within 3 days after the competition ends. The solution must reproduce the score on the public & private leaderboard.", "votes": null }, { "id": "1717133", "postDate": "03/09/2022 17:18:34", "content": "

How many is \"some sessions\" and do they have to be paid sessions?

", "rawMarkdown": "How many is \"some sessions\" and do they have to be paid sessions?", "votes": null }, { "id": "1717134", "postDate": "03/09/2022 17:19:27", "content": "

I am a completer newbie to Kaggle….where should I start to make the most out of this journey?

", "rawMarkdown": "I am a completer newbie to Kaggle....where should I start to make the most out of this journey?", "votes": null }, { "id": "1717184", "postDate": "03/09/2022 17:56:09", "content": "

at least 1 and nope.

", "rawMarkdown": "at least 1 and nope.", "votes": null }, { "id": "1717187", "postDate": "03/09/2022 17:56:57", "content": "

you can start from basic image classification problem. for example cats vs dogs but with 28 classes.

", "rawMarkdown": "you can start from basic image classification problem. for example cats vs dogs but with 28 classes.", "votes": null }, { "id": "1717377", "postDate": "03/09/2022 22:00:15", "content": "

I don't see it in the rules, but to confirm: are there no rules in terms of data augmentation? I.e., you can augment the images for training and for inference/testing?

", "rawMarkdown": "I don't see it in the rules, but to confirm: are there no rules in terms of data augmentation? I.e., you can augment the images for training and for inference/testing?", "votes": null }, { "id": "1718546", "postDate": "03/11/2022 00:24:38", "content": "

Hi @abhishek, Thank you for organizing the competition! Let me ask you two questions.

\n
    \n

  1. If I use a pretrained model with mnist, is it still eligible for prize? (In my humble opinion, it would be an interesting competition if this were explicitly prohibited.)

    2. \n

  2. You mentioned on the discord that the competition will last 1-2 weeks, but the timeline page on kaggle says it will last until April 10. Which is correct?

    3. \n
", "rawMarkdown": "Hi @abhishek, Thank you for organizing the competition! Let me ask you two questions.\n\n1. If I use a pretrained model with mnist, is it still eligible for prize? (In my humble opinion, it would be an interesting competition if this were explicitly prohibited.)\n\n2. You mentioned on the discord that the competition will last 1-2 weeks, but the timeline page on kaggle says it will last until April 10. Which is correct?", "votes": null }, { "id": "1719489", "postDate": "03/11/2022 20:03:17", "content": "

Thank you for organizing the competition!

", "rawMarkdown": "Thank you for organizing the competition!", "votes": null }, { "id": "1719846", "postDate": "03/12/2022 07:28:35", "content": "

Dear @err504

\n

…or perhaps even better still start with the basic image classification problem MNIST, a classic dataset of handwritten images with 10 classes.

\n

All the best,
\ncarl

", "rawMarkdown": "Dear @err504 \n\n...or perhaps even better still start with the basic image classification problem [MNIST, a classic dataset of handwritten images](https://www.kaggle.com/c/digit-recognizer) with 10 classes.\n\nAll the best,\ncarl", "votes": null }, { "id": "1720233", "postDate": "03/12/2022 15:33:40", "content": "

yes you can

", "rawMarkdown": "yes you can", "votes": null }, { "id": "1729025", "postDate": "03/19/2022 15:01:12", "content": "

Hi @abhishek, just want to clarify one thing that I have already registered for a few sessions, so do I need to sign up again to register through the link you provided, or can I just open this link, sign in to my existing account and register for few more sessions?

", "rawMarkdown": "Hi @abhishek, just want to clarify one thing that I have already registered for a few sessions, so do I need to sign up again to register through the link you provided, or can I just open this link, sign in to my existing account and register for few more sessions?", "votes": null }, { "id": "1730345", "postDate": "03/21/2022 06:50:36", "content": "

@abhishek

\n
\n

GTC sessions. Note: Prizes will be awarded only to those who register for GTC using the link above and attend some sessions.
\n Any session or GTC session because i am not able to find gtc category particularly.

\n
", "rawMarkdown": "abhishek \n> GTC sessions. Note: Prizes will be awarded only to those who register for GTC using the link above and attend some sessions.\nAny session or GTC session because i am not able to find gtc category particularly.", "votes": null } ], "comments": [ { "id": 1717133, "author_name": "nbroad", "author_url": "", "post_date": "03/09/2022 17:18:34", "content": "

How many is \"some sessions\" and do they have to be paid sessions?

", "votes": null, "replies": [ { "id": 1717184, "author_name": "abhishek", "author_url": "", "post_date": "03/09/2022 17:56:09", "content": "

at least 1 and nope.

", "votes": null, "replies": [] } ] }, { "id": 1717134, "author_name": "err504", "author_url": "", "post_date": "03/09/2022 17:19:27", "content": "

I am a completer newbie to Kaggle….where should I start to make the most out of this journey?

", "votes": null, "replies": [ { "id": 1717187, "author_name": "abhishek", "author_url": "", "post_date": "03/09/2022 17:56:57", "content": "

you can start from basic image classification problem. for example cats vs dogs but with 28 classes.

", "votes": null, "replies": [] }, { "id": 1719846, "author_name": "carlmcbrideellis", "author_url": "", "post_date": "03/12/2022 07:28:35", "content": "

Dear @err504

\n

…or perhaps even better still start with the basic image classification problem MNIST, a classic dataset of handwritten images with 10 classes.

\n

All the best,
\ncarl

", "votes": null, "replies": [] } ] }, { "id": 1717377, "author_name": "bramvanroy", "author_url": "", "post_date": "03/09/2022 22:00:15", "content": "

I don't see it in the rules, but to confirm: are there no rules in terms of data augmentation? I.e., you can augment the images for training and for inference/testing?

", "votes": null, "replies": [ { "id": 1720233, "author_name": "abhishek", "author_url": "", "post_date": "03/12/2022 15:33:40", "content": "

yes you can

", "votes": null, "replies": [] } ] }, { "id": 1718546, "author_name": "zaburo", "author_url": "", "post_date": "03/11/2022 00:24:38", "content": "

Hi @abhishek, Thank you for organizing the competition! Let me ask you two questions.

\n
    \n

  1. If I use a pretrained model with mnist, is it still eligible for prize? (In my humble opinion, it would be an interesting competition if this were explicitly prohibited.)

    2. \n

  2. You mentioned on the discord that the competition will last 1-2 weeks, but the timeline page on kaggle says it will last until April 10. Which is correct?

    3. \n
", "votes": null, "replies": [] }, { "id": 1719489, "author_name": "dmitrysergeevich", "author_url": "", "post_date": "03/11/2022 20:03:17", "content": "

Thank you for organizing the competition!

", "votes": null, "replies": [] }, { "id": 1729025, "author_name": "malayjoshi13", "author_url": "", "post_date": "03/19/2022 15:01:12", "content": "

Hi @abhishek, just want to clarify one thing that I have already registered for a few sessions, so do I need to sign up again to register through the link you provided, or can I just open this link, sign in to my existing account and register for few more sessions?

", "votes": null, "replies": [] }, { "id": 1730345, "author_name": "devanshchowdhury", "author_url": "", "post_date": "03/21/2022 06:50:36", "content": "

@abhishek

\n
\n

GTC sessions. Note: Prizes will be awarded only to those who register for GTC using the link above and attend some sessions.
\n Any session or GTC session because i am not able to find gtc category particularly.

\n
", "votes": null, "replies": [] } ], "raw_markdown_by_id": { "1716988": "In order to be eligible for the prizes, you must:\n\n- Register for NVIDIA GTC using this link: https://www.nvidia.com/gtc/?ncid=ref-crea-864532\n\n- Attend GTC sessions. Note: Prizes will be awarded only to those who register for GTC using the link above and attend some sessions.\n\n- Winners must open-source their solution within 3 days after the competition ends. The solution must reproduce the score on the public & private leaderboard.", "1717133": "How many is \"some sessions\" and do they have to be paid sessions?", "1717134": "I am a completer newbie to Kaggle....where should I start to make the most out of this journey?", "1717184": "at least 1 and nope.", "1717187": "you can start from basic image classification problem. for example cats vs dogs but with 28 classes.", "1717377": "I don't see it in the rules, but to confirm: are there no rules in terms of data augmentation? I.e., you can augment the images for training and for inference/testing?", "1718546": "Hi @abhishek, Thank you for organizing the competition! Let me ask you two questions.\n\n1. If I use a pretrained model with mnist, is it still eligible for prize? (In my humble opinion, it would be an interesting competition if this were explicitly prohibited.)\n\n2. You mentioned on the discord that the competition will last 1-2 weeks, but the timeline page on kaggle says it will last until April 10. Which is correct?", "1719489": "Thank you for organizing the competition!", "1719846": "Dear @err504 \n\n...or perhaps even better still start with the basic image classification problem [MNIST, a classic dataset of handwritten images](https://www.kaggle.com/c/digit-recognizer) with 10 classes.\n\nAll the best,\ncarl", "1720233": "yes you can", "1729025": "Hi @abhishek, just want to clarify one thing that I have already registered for a few sessions, so do I need to sign up again to register through the link you provided, or can I just open this link, sign in to my existing account and register for few more sessions?", "1730345": "abhishek \n> GTC sessions. Note: Prizes will be awarded only to those who register for GTC using the link above and attend some sessions.\nAny session or GTC session because i am not able to find gtc category particularly." }, "source": "meta" }