\n
![\"COTS\"](\"https://www.kaggleusercontent.com/kf/86551802/eyJhbGciOiJkaXIiLCJlbmMiOiJBMTI4Q0JDLUhTMjU2In0..lqdTbAEb7WWFta6kA4PoUg.Du-sWlx3daNLfBOnuFiDzFL7rlmesatVINvaI1LTaHZZpCiXzaT7GfjgqPr4CUMpnMGVpUGOvs7rlkbqJ4kJ5vOO4jzG_2aGq1g2v3yzOiWkV6TotVKr7ZVwqcRjV5KiRgp4AUzd9_6x3Bz4z_OZWtTYS3w7k7JM3M6VFsrteH-dz6qn52YNva5-HIud9EvgcEPhcxsNMMBBtZ-PFHe9_gOibMZIMiDs7QxrN4nZjHtX-g6jHpvF-3KwHvQk9SJZ5l4j9BEGTVRn8hxl1KsxzH10gpLanFquU0C3m8X7HHsC0DBuDr4r5A6CMTg8VgN9uzAnr_9fNIF2cZ4ESy5wb04N1nG-CzGQj6e-xRPznqd9HVEbAkeUbCOBEUVzbxK8O-7XkwMyGgys24pHeQIoaGj-XKfswk_xXG5X0RRgZmX5NM3raUwhQx4Y4p0BDSteOHVMmcsUetVGz_EScZ0CD6WN12kJfgz66hfz-phpW2aCxxvdId4is1r4svlA4ar6J3HFkzehYu1X_lZJgbdZ-iesy4UcnL1tKHGxvOpbWxRzlZz3AcSSBzWMn_obQgS-Mws2p1_mwDhmQHIsC8Qjo0CB1CGJ7QUBtWKSGyev-num-rAyOQlxyOld8jUpki3cLoERbMI1pg_G_obhEQNFF0vWbtVdApBc2j_xnoLAJ7BpK509vZ4gVmS3fSmQa3AA.qMfoDjS5aT0h6ny0f-rFzA/__results___files/__results___7_2.png\")
Dear Kagglers,
\nWe are exploring multiple options for how to augment the data. One interesting approach is to use GAN. Already trained network in action. When you will be interested, I consider publishing the training notebook.
\nGan Generator Notebook
\nI have also published the neural style transfer augmentation method that you can check it also
\nNeural Style Transfer Notebook
Do you need 10, 100, or 1M starfishes? No problem.
\n
Dear Kagglers,
\nWe are exploring multiple options for how to augment the data. One interesting approach is to use GAN. Already trained network in action. When you will be interested, I consider publishing the training notebook.
\nGan Generator Notebook
\nI have also published the neural style transfer augmentation method that you can check it also
\nNeural Style Transfer Notebook
Do you need 10, 100, or 1M starfishes? No problem.
\n
If you need the model:
\nModel Dataset
Thank you for sharing ideas.❤️
", "rawMarkdown": "Thank you for sharing ideas.❤️", "votes": null }, { "id": "1669405", "postDate": "01/30/2022 16:38:31", "content": "Thank you @locbaop :)
", "rawMarkdown": "Thank you @locbaop :)", "votes": null }, { "id": "1669450", "postDate": "01/30/2022 17:33:57", "content": "So cool! Thank you for sharing (and glad COTS don't get generated as easily in real life 😁).
", "rawMarkdown": "So cool! Thank you for sharing (and glad COTS don't get generated as easily in real life 😁).", "votes": null }, { "id": "1669454", "postDate": "01/30/2022 17:39:01", "content": "Thank you @dauriel not that easy as GAN but unfortunately there are too many cots in the seas.
", "rawMarkdown": "Thank you @dauriel not that easy as GAN but unfortunately there are too many cots in the seas.", "votes": null }, { "id": "1669803", "postDate": "01/31/2022 00:37:46", "content": "Thank you for sharing!
", "rawMarkdown": "Thank you for sharing!", "votes": null }, { "id": "1669898", "postDate": "01/31/2022 03:45:12", "content": "how long did origial gan model train?
", "rawMarkdown": "how long did origial gan model train?", "votes": null }, { "id": "1669933", "postDate": "01/31/2022 04:49:04", "content": "Google colab pro - 5h :)
", "rawMarkdown": "Google colab pro - 5h :)", "votes": null }, { "id": "1669934", "postDate": "01/31/2022 04:49:47", "content": "You are welcome @brendanrappazzo :)
", "rawMarkdown": "You are welcome @brendanrappazzo :)", "votes": null }, { "id": "1670109", "postDate": "01/31/2022 08:23:48", "content": "Nice work. GANs are definitely my favorite genre of machine learning models. A machine with creativity even after years I'm still in awe when it works.
", "rawMarkdown": "Nice work. GANs are definitely my favorite genre of machine learning models. A machine with creativity even after years I'm still in awe when it works.", "votes": null }, { "id": "1670128", "postDate": "01/31/2022 08:38:09", "content": "Fully agree @taranmarley! A lot of fun with GANS but training competitive networks in the same time is challenging :)
", "rawMarkdown": "Fully agree @taranmarley! A lot of fun with GANS but training competitive networks in the same time is challenging :)", "votes": null }, { "id": "1670279", "postDate": "01/31/2022 11:36:57", "content": "Great work, thanks for sharing. Did you trained the GAN using this competition's dataset? Or did you also use some external datasets?
", "rawMarkdown": "Great work, thanks for sharing. Did you trained the GAN using this competition's dataset? Or did you also use some external datasets?", "votes": null }, { "id": "1670450", "postDate": "01/31/2022 14:36:54", "content": "Thank you @kfk42kfk I have used the training dataset. If the people will be interested in GAN solution I will prepare and publish the full ETL and Training :)
", "rawMarkdown": "Thank you @kfk42kfk I have used the training dataset. If the people will be interested in GAN solution I will prepare and publish the full ETL and Training :)", "votes": null }, { "id": "1674250", "postDate": "02/03/2022 10:49:21", "content": "You can find GAN training pipeline here: GAN Training Pipeline. I have recently published the notebook :)
", "rawMarkdown": "You can find GAN training pipeline here: [GAN Training Pipeline](https://www.kaggle.com/marcinstasko/gan-training-make-unlimited-cots). I have recently published the notebook :)", "votes": null }, { "id": "1682874", "postDate": "02/09/2022 12:38:45", "content": "Tell me In order to use Tensorflow without interruptions is it nessecary to change to TPU from GPU on our Notebooks.
", "rawMarkdown": "Tell me In order to use Tensorflow without interruptions is it nessecary to change to TPU from GPU on our Notebooks.", "votes": null }, { "id": "1682927", "postDate": "02/09/2022 13:22:54", "content": "Dear @muhammadammarjamshed I am using Pytorch on this notebook :)
", "rawMarkdown": "Dear @muhammadammarjamshed I am using Pytorch on this notebook :)", "votes": null }, { "id": "1683181", "postDate": "02/09/2022 16:14:39", "content": "Yes Thankyou for clarifying that but did you run the code on GPU or TPU !
", "rawMarkdown": "Yes Thankyou for clarifying that but did you run the code on GPU or TPU !", "votes": null } ], "comments": [ { "id": 1669314, "author_name": "marcinstasko", "author_url": "", "post_date": "01/30/2022 15:14:51", "content": "If you need the model:
\nModel Dataset
Thank you for sharing ideas.❤️
", "votes": null, "replies": [ { "id": 1669405, "author_name": "marcinstasko", "author_url": "", "post_date": "01/30/2022 16:38:31", "content": "Thank you @locbaop :)
", "votes": null, "replies": [] } ] }, { "id": 1669450, "author_name": "dauriel", "author_url": "", "post_date": "01/30/2022 17:33:57", "content": "So cool! Thank you for sharing (and glad COTS don't get generated as easily in real life 😁).
", "votes": null, "replies": [ { "id": 1669454, "author_name": "marcinstasko", "author_url": "", "post_date": "01/30/2022 17:39:01", "content": "Thank you @dauriel not that easy as GAN but unfortunately there are too many cots in the seas.
", "votes": null, "replies": [] } ] }, { "id": 1669803, "author_name": "brendanrappazzo", "author_url": "", "post_date": "01/31/2022 00:37:46", "content": "Thank you for sharing!
", "votes": null, "replies": [ { "id": 1669934, "author_name": "marcinstasko", "author_url": "", "post_date": "01/31/2022 04:49:47", "content": "You are welcome @brendanrappazzo :)
", "votes": null, "replies": [] } ] }, { "id": 1669898, "author_name": "drzhuzhe", "author_url": "", "post_date": "01/31/2022 03:45:12", "content": "how long did origial gan model train?
", "votes": null, "replies": [ { "id": 1669933, "author_name": "marcinstasko", "author_url": "", "post_date": "01/31/2022 04:49:04", "content": "Google colab pro - 5h :)
", "votes": null, "replies": [] } ] }, { "id": 1670109, "author_name": "taranmarley", "author_url": "", "post_date": "01/31/2022 08:23:48", "content": "Nice work. GANs are definitely my favorite genre of machine learning models. A machine with creativity even after years I'm still in awe when it works.
", "votes": null, "replies": [ { "id": 1670128, "author_name": "marcinstasko", "author_url": "", "post_date": "01/31/2022 08:38:09", "content": "Fully agree @taranmarley! A lot of fun with GANS but training competitive networks in the same time is challenging :)
", "votes": null, "replies": [] } ] }, { "id": 1670279, "author_name": "kfk42kfk", "author_url": "", "post_date": "01/31/2022 11:36:57", "content": "Great work, thanks for sharing. Did you trained the GAN using this competition's dataset? Or did you also use some external datasets?
", "votes": null, "replies": [ { "id": 1670450, "author_name": "marcinstasko", "author_url": "", "post_date": "01/31/2022 14:36:54", "content": "Thank you @kfk42kfk I have used the training dataset. If the people will be interested in GAN solution I will prepare and publish the full ETL and Training :)
", "votes": null, "replies": [] } ] }, { "id": 1674250, "author_name": "marcinstasko", "author_url": "", "post_date": "02/03/2022 10:49:21", "content": "You can find GAN training pipeline here: GAN Training Pipeline. I have recently published the notebook :)
", "votes": null, "replies": [] }, { "id": 1682874, "author_name": "muhammadammarjamshed", "author_url": "", "post_date": "02/09/2022 12:38:45", "content": "Tell me In order to use Tensorflow without interruptions is it nessecary to change to TPU from GPU on our Notebooks.
", "votes": null, "replies": [ { "id": 1682927, "author_name": "marcinstasko", "author_url": "", "post_date": "02/09/2022 13:22:54", "content": "Dear @muhammadammarjamshed I am using Pytorch on this notebook :)
", "votes": null, "replies": [] }, { "id": 1683181, "author_name": "muhammadammarjamshed", "author_url": "", "post_date": "02/09/2022 16:14:39", "content": "Yes Thankyou for clarifying that but did you run the code on GPU or TPU !
", "votes": null, "replies": [] } ] } ], "raw_markdown_by_id": { "1669200": "Dear Kagglers,\n\nWe are exploring multiple options for how to augment the data. One interesting approach is to use GAN. Already trained network in action. When you will be interested, I consider publishing the training notebook.\n[Gan Generator Notebook](https://www.kaggle.com/marcinstasko/unlimited-cots-generator-pytorch-gan-in-action)\nI have also published the neural style transfer augmentation method that you can check it also \n[Neural Style Transfer Notebook](https://www.kaggle.com/marcinstasko/cots-neuralstyle-transfer-pytorch-augumentation)\n\nDo you need 10, 100, or 1M starfishes? No problem.\n", "1669314": "If you need the model:\n[Model Dataset](https://www.kaggle.com/marcinstasko/cots-ganv1)", "1669383": "Thank you for sharing ideas.❤️", "1669405": "Thank you @locbaop :)", "1669450": "So cool! Thank you for sharing (and glad COTS don't get generated as easily in real life 😁).", "1669454": "Thank you @dauriel not that easy as GAN but unfortunately there are too many cots in the seas.", "1669803": "Thank you for sharing!", "1669898": "how long did origial gan model train?", "1669933": "Google colab pro - 5h :)", "1669934": "You are welcome @brendanrappazzo :)", "1670109": "Nice work. GANs are definitely my favorite genre of machine learning models. A machine with creativity even after years I'm still in awe when it works.", "1670128": "Fully agree @taranmarley! A lot of fun with GANS but training competitive networks in the same time is challenging :)", "1670279": "Great work, thanks for sharing. Did you trained the GAN using this competition's dataset? Or did you also use some external datasets?", "1670450": "Thank you @kfk42kfk I have used the training dataset. If the people will be interested in GAN solution I will prepare and publish the full ETL and Training :)", "1674250": "You can find GAN training pipeline here: [GAN Training Pipeline](https://www.kaggle.com/marcinstasko/gan-training-make-unlimited-cots). I have recently published the notebook :)", "1682874": "Tell me In order to use Tensorflow without interruptions is it nessecary to change to TPU from GPU on our Notebooks.", "1682927": "Dear @muhammadammarjamshed I am using Pytorch on this notebook :)", "1683181": "Yes Thankyou for clarifying that but did you run the code on GPU or TPU !" }, "source": "meta" }