I'm excited to announce that, after an evaluation of contributions to this competition, we have identified the following 2 newest TPU Stars, awarded to the most knowledgeable and helpful TPU experts in this competition.
\n\n✨ xhlulu, who was the first to provide an excellent, easy to read guide for using XLM-Roberta with TensorFlow in Kaggle notebooks! \n✨ Alex Shonenkov who shared well-documented and helpful training and inference notebooks, and provided great support in addressing some of the pain points everyone was facing in using Pytorch on TPUs, including this notebook demonstrating how to maintain class balance during distributed TPU training.
\n\nThanks to both of you and the many of the other peer-acknowledged participants for your contributions and sharing to support one another in learning how to use TPUs!
", "messages": [ { "id": "952327", "postDate": "07/30/2020 20:21:17", "content": "I'm excited to announce that, after an evaluation of contributions to this competition, we have identified the following 2 newest TPU Stars, awarded to the most knowledgeable and helpful TPU experts in this competition.
\n\n✨ xhlulu, who was the first to provide an excellent, easy to read guide for using XLM-Roberta with TensorFlow in Kaggle notebooks! \n✨ Alex Shonenkov who shared well-documented and helpful training and inference notebooks, and provided great support in addressing some of the pain points everyone was facing in using Pytorch on TPUs, including this notebook demonstrating how to maintain class balance during distributed TPU training.
\n\nThanks to both of you and the many of the other peer-acknowledged participants for your contributions and sharing to support one another in learning how to use TPUs!
", "rawMarkdown": "I'm excited to announce that, after an evaluation of contributions to this competition, we have identified the following 2 newest TPU Stars, awarded to the most knowledgeable and helpful TPU experts in this competition.\n\n✨ **[xhlulu](https://www.kaggle.com/xhlulu)**, who was the first to provide an excellent, easy to read guide for using [XLM-Roberta](https://www.kaggle.com/xhlulu/jigsaw-tpu-xlm-roberta) with TensorFlow in Kaggle notebooks! \n✨ **[Alex Shonenkov](https://www.kaggle.com/shonenkov)** who shared well-documented and helpful [training](https://www.kaggle.com/shonenkov/tpu-training-super-fast-xlmroberta/data) and [inference](https://www.kaggle.com/shonenkov/tpu-inference-super-fast-xlmroberta/data) notebooks, and provided great support in addressing some of the pain points everyone was facing in using Pytorch on TPUs, including [this notebook](https://www.kaggle.com/shonenkov/class-balance-with-pytorch-xla) demonstrating how to maintain class balance during distributed TPU training.\n\nThanks to both of you and [the many of the other peer-acknowledged participants](https://www.kaggle.com/c/jigsaw-multilingual-toxic-comment-classification/discussion/160844) for your contributions and sharing to support one another in learning how to use TPUs!", "votes": null }, { "id": "952766", "postDate": "07/31/2020 07:51:31", "content": "good nice to hear
", "rawMarkdown": "good nice to hear", "votes": null }, { "id": "953178", "postDate": "07/31/2020 15:26:33", "content": "@juliaelliott How about other competition? E.g. ALASKA2 Image Steganalysis
", "rawMarkdown": "juliaelliott How about other competition? E.g. [ALASKA2 Image Steganalysis](https://www.kaggle.com/c/alaska2-image-steganalysis)", "votes": null }, { "id": "953182", "postDate": "07/31/2020 15:29:28", "content": "If you read the prize section of Toxic Comp here, you will see that they decided to award $10,000 to TPU stars. Alaska didn't have prizes for TPU stars.
", "rawMarkdown": "If you read the prize section of Toxic Comp [here][1], you will see that they decided to award $10,000 to TPU stars. Alaska didn't have prizes for TPU stars.\n\n[1]: https://www.kaggle.com/c/jigsaw-multilingual-toxic-comment-classification/overview/prizes", "votes": null }, { "id": "953187", "postDate": "07/31/2020 15:34:50", "content": "@cdeotte Thank you. But I found this page https://www.kaggle.com/tpu-prize
", "rawMarkdown": "cdeotte Thank you. But I found this page [https://www.kaggle.com/tpu-prize](https://www.kaggle.com/tpu-prize)", "votes": null }, { "id": "953381", "postDate": "07/31/2020 18:28:57", "content": "@wuliaokaola That link you shared is a separate general TPU program that is not specific to a competition. To be eligible, you need to meet the criteria specified on that page and submit through that page's link. Candidates are evaluated and awarded on a rolling basis.
", "rawMarkdown": "wuliaokaola That link you shared is a separate general TPU program that is not specific to a competition. To be eligible, you need to meet the criteria specified on that page and submit through that page's link. Candidates are evaluated and awarded on a rolling basis.", "votes": null }, { "id": "953650", "postDate": "08/01/2020 00:37:08", "content": "@juliaelliott Thank you. I submitted a notebook of mine. https://www.kaggle.com/wuliaokaola/alaska2-some-tips-for-using-tpu
", "rawMarkdown": "juliaelliott Thank you. I submitted a notebook of mine. https://www.kaggle.com/wuliaokaola/alaska2-some-tips-for-using-tpu", "votes": null }, { "id": "953660", "postDate": "08/01/2020 01:05:59", "content": "Thanks @wuliaokaola for letting me know about this. I submitted a notebook too.
", "rawMarkdown": "Thanks @wuliaokaola for letting me know about this. I submitted a notebook too.", "votes": null }, { "id": "953896", "postDate": "08/01/2020 08:18:01", "content": "Congratulations to the winners!
", "rawMarkdown": "Congratulations to the winners!", "votes": null }, { "id": "953907", "postDate": "08/01/2020 08:31:43", "content": "Congrats @xhlulu @shonenkov ! Thanks for sharing helpful discussions and notebooks.
", "rawMarkdown": "Congrats @xhlulu @shonenkov ! Thanks for sharing helpful discussions and notebooks.", "votes": null }, { "id": "956276", "postDate": "08/03/2020 11:38:55", "content": "Congrats @xhlulu and @shonenkov for your contributions
", "rawMarkdown": "Congrats @xhlulu and @shonenkov for your contributions", "votes": null }, { "id": "956931", "postDate": "08/03/2020 23:40:11", "content": "Congratulations @xhlulu and @shonenkov
", "rawMarkdown": "Congratulations @xhlulu and @shonenkov", "votes": null }, { "id": "958097", "postDate": "08/04/2020 19:02:52", "content": "Congrats Everyone !!
", "rawMarkdown": "Congrats Everyone !!", "votes": null }, { "id": "962435", "postDate": "08/08/2020 06:02:59", "content": "Nice to hear!
", "rawMarkdown": "Nice to hear!", "votes": null } ], "comments": [ { "id": 952766, "author_name": "likhithkvittla", "author_url": "", "post_date": "07/31/2020 07:51:31", "content": "good nice to hear
", "votes": null, "replies": [] }, { "id": 953178, "author_name": "wuliaokaola", "author_url": "", "post_date": "07/31/2020 15:26:33", "content": "@juliaelliott How about other competition? E.g. ALASKA2 Image Steganalysis
", "votes": null, "replies": [ { "id": 953182, "author_name": "cdeotte", "author_url": "", "post_date": "07/31/2020 15:29:28", "content": "If you read the prize section of Toxic Comp here, you will see that they decided to award $10,000 to TPU stars. Alaska didn't have prizes for TPU stars.
", "votes": null, "replies": [] }, { "id": 953187, "author_name": "wuliaokaola", "author_url": "", "post_date": "07/31/2020 15:34:50", "content": "@cdeotte Thank you. But I found this page https://www.kaggle.com/tpu-prize
", "votes": null, "replies": [] }, { "id": 953381, "author_name": "juliaelliott", "author_url": "", "post_date": "07/31/2020 18:28:57", "content": "@wuliaokaola That link you shared is a separate general TPU program that is not specific to a competition. To be eligible, you need to meet the criteria specified on that page and submit through that page's link. Candidates are evaluated and awarded on a rolling basis.
", "votes": null, "replies": [] }, { "id": 953650, "author_name": "wuliaokaola", "author_url": "", "post_date": "08/01/2020 00:37:08", "content": "@juliaelliott Thank you. I submitted a notebook of mine. https://www.kaggle.com/wuliaokaola/alaska2-some-tips-for-using-tpu
", "votes": null, "replies": [] }, { "id": 953660, "author_name": "cdeotte", "author_url": "", "post_date": "08/01/2020 01:05:59", "content": "Thanks @wuliaokaola for letting me know about this. I submitted a notebook too.
", "votes": null, "replies": [] } ] }, { "id": 953896, "author_name": "riblidezso", "author_url": "", "post_date": "08/01/2020 08:18:01", "content": "Congratulations to the winners!
", "votes": null, "replies": [] }, { "id": 953907, "author_name": "cdeotte", "author_url": "", "post_date": "08/01/2020 08:31:43", "content": "Congrats @xhlulu @shonenkov ! Thanks for sharing helpful discussions and notebooks.
", "votes": null, "replies": [] }, { "id": 956276, "author_name": "vijayakishoredusi", "author_url": "", "post_date": "08/03/2020 11:38:55", "content": "Congrats @xhlulu and @shonenkov for your contributions
", "votes": null, "replies": [] }, { "id": 956931, "author_name": "salmaneunus", "author_url": "", "post_date": "08/03/2020 23:40:11", "content": "Congratulations @xhlulu and @shonenkov
", "votes": null, "replies": [] }, { "id": 958097, "author_name": "shaitender", "author_url": "", "post_date": "08/04/2020 19:02:52", "content": "Congrats Everyone !!
", "votes": null, "replies": [] }, { "id": 962435, "author_name": "shahad001", "author_url": "", "post_date": "08/08/2020 06:02:59", "content": "Nice to hear!
", "votes": null, "replies": [] } ], "raw_markdown_by_id": { "952327": "I'm excited to announce that, after an evaluation of contributions to this competition, we have identified the following 2 newest TPU Stars, awarded to the most knowledgeable and helpful TPU experts in this competition.\n\n✨ **[xhlulu](https://www.kaggle.com/xhlulu)**, who was the first to provide an excellent, easy to read guide for using [XLM-Roberta](https://www.kaggle.com/xhlulu/jigsaw-tpu-xlm-roberta) with TensorFlow in Kaggle notebooks! \n✨ **[Alex Shonenkov](https://www.kaggle.com/shonenkov)** who shared well-documented and helpful [training](https://www.kaggle.com/shonenkov/tpu-training-super-fast-xlmroberta/data) and [inference](https://www.kaggle.com/shonenkov/tpu-inference-super-fast-xlmroberta/data) notebooks, and provided great support in addressing some of the pain points everyone was facing in using Pytorch on TPUs, including [this notebook](https://www.kaggle.com/shonenkov/class-balance-with-pytorch-xla) demonstrating how to maintain class balance during distributed TPU training.\n\nThanks to both of you and [the many of the other peer-acknowledged participants](https://www.kaggle.com/c/jigsaw-multilingual-toxic-comment-classification/discussion/160844) for your contributions and sharing to support one another in learning how to use TPUs!", "952766": "good nice to hear", "953178": "juliaelliott How about other competition? E.g. [ALASKA2 Image Steganalysis](https://www.kaggle.com/c/alaska2-image-steganalysis)", "953182": "If you read the prize section of Toxic Comp [here][1], you will see that they decided to award $10,000 to TPU stars. Alaska didn't have prizes for TPU stars.\n\n[1]: https://www.kaggle.com/c/jigsaw-multilingual-toxic-comment-classification/overview/prizes", "953187": "cdeotte Thank you. But I found this page [https://www.kaggle.com/tpu-prize](https://www.kaggle.com/tpu-prize)", "953381": "wuliaokaola That link you shared is a separate general TPU program that is not specific to a competition. To be eligible, you need to meet the criteria specified on that page and submit through that page's link. Candidates are evaluated and awarded on a rolling basis.", "953650": "juliaelliott Thank you. I submitted a notebook of mine. https://www.kaggle.com/wuliaokaola/alaska2-some-tips-for-using-tpu", "953660": "Thanks @wuliaokaola for letting me know about this. I submitted a notebook too.", "953896": "Congratulations to the winners!", "953907": "Congrats @xhlulu @shonenkov ! Thanks for sharing helpful discussions and notebooks.", "956276": "Congrats @xhlulu and @shonenkov for your contributions", "956931": "Congratulations @xhlulu and @shonenkov", "958097": "Congrats Everyone !!", "962435": "Nice to hear!" }, "source": "meta" }