{ "id": 349991, "title": "Winners Announcement and Final Leaderboard!!", "url": "/competitions/dlsprint/discussion/349991", "author_name": "", "post_date": "2022-09-03T17:49:36.827661900Z", "votes": 5, "comment_count": 4, "views": 0, "content": "

After two whole months, the competition finally came to a close on September 2, 2022. Here are the winners by categories:

\n
\n

Best Notebooks:

\n
    \n
  1. @shahruk10 for his notebook : Error Analysis using SCTK
    2. \n
  2. @mbmmurad for his notebook: End to end Pipeline - CNN+RNN model with CTC loss
    3. \n
  3. @shourovj for his notebook: Preprocessing and Dataset Class and Tokenizer
    4. \n
\n
\n

BUET Rising Team:

\n

Team YellowKing
\nMembers: @sameen53, @salmankhondker, @somonindroroy, @arcfahad

\n
\n

Final Results:

\n

Champions

\n

Team YellowKing
\nMembers: @sameen53, @salmankhondker, @somonindroroy, @arcfahad

\n

1st Runners Up

\n

Team একুশ
\nMembers: @shahruk10, @sknahin, @sanjayacharjee, @sawradipsaha

\n

2nd Runners Up

\n

Team Bucephalus NSU
\nMembers: @mohammedrakib, @ismail31416

\n
\n

The Final leaderboard is given here:
\n\"\"

\n

The final leaderboard is based on Hidden dataset score, inference time, Kaggle private leaderboard score and Judge Marks. We thank everyone for being with us for so long and hope you guys enjoyed this competiton. We hope through this competition we have taken the Bengali ASR research to new heights. Please stay tuned for more ;)

\n
\n

Call for Papers:

\n

We are writing a paper on this competition. We would like to cite each and every one of your works in this paper. That is why we are instructing every one of the teams to revamp your papers and submit to arxiv and send us the link at dlsprint2022@gmail.com. Please use your team name as the email subject. The deadline is September 11, 2022. In order to expedite this process, we are available for help and advice. Please give edit access to sushmit@ieee.org to your overleaf/google doc. We won't be a part of the author list, we want to give feedback, expedite the process and help you in any way we can.

\n

Thank you everyone. Till next time. Good luck.

", "messages": [ { "id": "1925203", "postDate": "09/03/2022 17:49:36", "content": "

After two whole months, the competition finally came to a close on September 2, 2022. Here are the winners by categories:

\n
\n

Best Notebooks:

\n
    \n
  1. @shahruk10 for his notebook : Error Analysis using SCTK
    2. \n
  2. @mbmmurad for his notebook: End to end Pipeline - CNN+RNN model with CTC loss
    3. \n
  3. @shourovj for his notebook: Preprocessing and Dataset Class and Tokenizer
    4. \n
\n
\n

BUET Rising Team:

\n

Team YellowKing
\nMembers: @sameen53, @salmankhondker, @somonindroroy, @arcfahad

\n
\n

Final Results:

\n

Champions

\n

Team YellowKing
\nMembers: @sameen53, @salmankhondker, @somonindroroy, @arcfahad

\n

1st Runners Up

\n

Team একুশ
\nMembers: @shahruk10, @sknahin, @sanjayacharjee, @sawradipsaha

\n

2nd Runners Up

\n

Team Bucephalus NSU
\nMembers: @mohammedrakib, @ismail31416

\n
\n

The Final leaderboard is given here:
\n\"\"

\n

The final leaderboard is based on Hidden dataset score, inference time, Kaggle private leaderboard score and Judge Marks. We thank everyone for being with us for so long and hope you guys enjoyed this competiton. We hope through this competition we have taken the Bengali ASR research to new heights. Please stay tuned for more ;)

\n
\n

Call for Papers:

\n

We are writing a paper on this competition. We would like to cite each and every one of your works in this paper. That is why we are instructing every one of the teams to revamp your papers and submit to arxiv and send us the link at dlsprint2022@gmail.com. Please use your team name as the email subject. The deadline is September 11, 2022. In order to expedite this process, we are available for help and advice. Please give edit access to sushmit@ieee.org to your overleaf/google doc. We won't be a part of the author list, we want to give feedback, expedite the process and help you in any way we can.

\n

Thank you everyone. Till next time. Good luck.

", "rawMarkdown": "After two whole months, the competition finally came to a close on September 2, 2022. Here are the winners by categories:\n\n------------------------------------------------------------------------------------------------------\n### Best Notebooks:\n1. @shahruk10 for his notebook : [Error Analysis using SCTK](https://www.kaggle.com/code/shahruk10/error-analysis-using-sctk)\n2. @mbmmurad for his notebook: [End to end Pipeline - CNN+RNN model with CTC loss](https://www.kaggle.com/code/mbmmurad/end-to-end-pipeline-cnn-rnn-model-with-ctc-loss)\n3. @shourovj for his notebook: [Preprocessing and Dataset Class and Tokenizer](https://www.kaggle.com/code/shourovj/preprocessing-and-dataset-class-and-tokenizer)\n\n------------------------------------------------------------------------------------------------------\n\n### BUET Rising Team: \nTeam YellowKing\nMembers: @sameen53, @salmankhondker, @somonindroroy, @arcfahad\n\n------------------------------------------------------------------------------------------------------\n# Final Results:\n### Champions\nTeam YellowKing\nMembers: @sameen53, @salmankhondker, @somonindroroy, @arcfahad\n\n### 1st Runners Up\nTeam একুশ\nMembers: @shahruk10, @sknahin, @sanjayacharjee, @sawradipsaha\n\n### 2nd Runners Up\nTeam Bucephalus NSU\nMembers: @mohammedrakib, @ismail31416\n\n------------------------------------------------------------------------------------------------------\nThe Final leaderboard is given here:\n![](https://www.googleapis.com/download/storage/v1/b/kaggle-forum-message-attachments/o/inbox%2F1530576%2F01184b4c1e5105e4b820336f970be3ba%2Ffinal_leaderboard.jpg?generation=1662226556754094&alt=media)\n\nThe final leaderboard is based on Hidden dataset score, inference time, Kaggle private leaderboard score and Judge Marks. We thank everyone for being with us for so long and hope you guys enjoyed this competiton. We hope through this competition we have taken the Bengali ASR research to new heights. Please stay tuned for more ;)\n\n------------------------------------------------------------------------------------------------------\n\n# Call for Papers:\nWe are writing a paper on this competition. We would like to cite each and every one of your works in this paper. That is why we are instructing every one of the teams to revamp your papers and submit to arxiv and send us the link at **dlsprint2022@gmail.com**. Please use your team name as the email subject. The deadline is **September 11, 2022**. In order to expedite this process, we are available for help and advice. Please give edit access to **sushmit@ieee.org** to your overleaf/google doc. We won't be a part of the author list, we want to give feedback, expedite the process and help you in any way we can.\n\nThank you everyone. Till next time. Good luck.", "votes": null }, { "id": "1925400", "postDate": "09/03/2022 23:27:13", "content": "

A minor update to the teams: please use this mail for the paper feedback (if you need any): sushmit0109@gmail.com. Thanks!

", "rawMarkdown": "A minor update to the teams: please use this mail for the paper feedback (if you need any): sushmit0109@gmail.com. Thanks!", "votes": null }, { "id": "1925651", "postDate": "09/04/2022 07:08:23", "content": "

To the competition host, are you planning to make the hidden dataset public as well? I guess It will help us in the error analysis of our models.

", "rawMarkdown": "To the competition host, are you planning to make the hidden dataset public as well? I guess It will help us in the error analysis of our models.", "votes": null }, { "id": "1926513", "postDate": "09/04/2022 22:25:37", "content": "

No, the dataset will remain private and serve as a benchmark for future Bengali ASR models too. If you want to do some more inferences on the testing data for your paper, you can share the kaggle inference notebook with me (like before) and I will provide you the scores. (Max: 3 inferences per team). You may reach out : sushmit@ieee.org

", "rawMarkdown": "No, the dataset will remain private and serve as a benchmark for future Bengali ASR models too. If you want to do some more inferences on the testing data for your paper, you can share the kaggle inference notebook with me (like before) and I will provide you the scores. (Max: 3 inferences per team). You may reach out : sushmit@ieee.org", "votes": null }, { "id": "1926515", "postDate": "09/04/2022 22:26:15", "content": "

Dear all, the dataset will remain private and serve as a benchmark for future Bengali ASR models too. If you want to do some more inferences on the testing data for your paper, you can share the kaggle inference notebook with me (like before) and I will provide you the scores. (Max: 3 inferences per team). You may reach out : sushmit@ieee.org

", "rawMarkdown": "Dear all, the dataset will remain private and serve as a benchmark for future Bengali ASR models too. If you want to do some more inferences on the testing data for your paper, you can share the kaggle inference notebook with me (like before) and I will provide you the scores. (Max: 3 inferences per team). You may reach out : sushmit@ieee.org", "votes": null } ], "comments": [ { "id": 1925400, "author_name": "sushmit0109", "author_url": "", "post_date": "09/03/2022 23:27:13", "content": "

A minor update to the teams: please use this mail for the paper feedback (if you need any): sushmit0109@gmail.com. Thanks!

", "votes": null, "replies": [] }, { "id": 1925651, "author_name": "afiaibnath", "author_url": "", "post_date": "09/04/2022 07:08:23", "content": "

To the competition host, are you planning to make the hidden dataset public as well? I guess It will help us in the error analysis of our models.

", "votes": null, "replies": [ { "id": 1926513, "author_name": "sushmit0109", "author_url": "", "post_date": "09/04/2022 22:25:37", "content": "

No, the dataset will remain private and serve as a benchmark for future Bengali ASR models too. If you want to do some more inferences on the testing data for your paper, you can share the kaggle inference notebook with me (like before) and I will provide you the scores. (Max: 3 inferences per team). You may reach out : sushmit@ieee.org

", "votes": null, "replies": [] } ] }, { "id": 1926515, "author_name": "sushmit0109", "author_url": "", "post_date": "09/04/2022 22:26:15", "content": "

Dear all, the dataset will remain private and serve as a benchmark for future Bengali ASR models too. If you want to do some more inferences on the testing data for your paper, you can share the kaggle inference notebook with me (like before) and I will provide you the scores. (Max: 3 inferences per team). You may reach out : sushmit@ieee.org

", "votes": null, "replies": [] } ], "raw_markdown_by_id": { "1925203": "After two whole months, the competition finally came to a close on September 2, 2022. Here are the winners by categories:\n\n------------------------------------------------------------------------------------------------------\n### Best Notebooks:\n1. @shahruk10 for his notebook : [Error Analysis using SCTK](https://www.kaggle.com/code/shahruk10/error-analysis-using-sctk)\n2. @mbmmurad for his notebook: [End to end Pipeline - CNN+RNN model with CTC loss](https://www.kaggle.com/code/mbmmurad/end-to-end-pipeline-cnn-rnn-model-with-ctc-loss)\n3. @shourovj for his notebook: [Preprocessing and Dataset Class and Tokenizer](https://www.kaggle.com/code/shourovj/preprocessing-and-dataset-class-and-tokenizer)\n\n------------------------------------------------------------------------------------------------------\n\n### BUET Rising Team: \nTeam YellowKing\nMembers: @sameen53, @salmankhondker, @somonindroroy, @arcfahad\n\n------------------------------------------------------------------------------------------------------\n# Final Results:\n### Champions\nTeam YellowKing\nMembers: @sameen53, @salmankhondker, @somonindroroy, @arcfahad\n\n### 1st Runners Up\nTeam একুশ\nMembers: @shahruk10, @sknahin, @sanjayacharjee, @sawradipsaha\n\n### 2nd Runners Up\nTeam Bucephalus NSU\nMembers: @mohammedrakib, @ismail31416\n\n------------------------------------------------------------------------------------------------------\nThe Final leaderboard is given here:\n![](https://www.googleapis.com/download/storage/v1/b/kaggle-forum-message-attachments/o/inbox%2F1530576%2F01184b4c1e5105e4b820336f970be3ba%2Ffinal_leaderboard.jpg?generation=1662226556754094&alt=media)\n\nThe final leaderboard is based on Hidden dataset score, inference time, Kaggle private leaderboard score and Judge Marks. We thank everyone for being with us for so long and hope you guys enjoyed this competiton. We hope through this competition we have taken the Bengali ASR research to new heights. Please stay tuned for more ;)\n\n------------------------------------------------------------------------------------------------------\n\n# Call for Papers:\nWe are writing a paper on this competition. We would like to cite each and every one of your works in this paper. That is why we are instructing every one of the teams to revamp your papers and submit to arxiv and send us the link at **dlsprint2022@gmail.com**. Please use your team name as the email subject. The deadline is **September 11, 2022**. In order to expedite this process, we are available for help and advice. Please give edit access to **sushmit@ieee.org** to your overleaf/google doc. We won't be a part of the author list, we want to give feedback, expedite the process and help you in any way we can.\n\nThank you everyone. Till next time. Good luck.", "1925400": "A minor update to the teams: please use this mail for the paper feedback (if you need any): sushmit0109@gmail.com. Thanks!", "1925651": "To the competition host, are you planning to make the hidden dataset public as well? I guess It will help us in the error analysis of our models.", "1926513": "No, the dataset will remain private and serve as a benchmark for future Bengali ASR models too. If you want to do some more inferences on the testing data for your paper, you can share the kaggle inference notebook with me (like before) and I will provide you the scores. (Max: 3 inferences per team). You may reach out : sushmit@ieee.org", "1926515": "Dear all, the dataset will remain private and serve as a benchmark for future Bengali ASR models too. If you want to do some more inferences on the testing data for your paper, you can share the kaggle inference notebook with me (like before) and I will provide you the scores. (Max: 3 inferences per team). You may reach out : sushmit@ieee.org" }, "source": "meta" }