{ "id": 154322, "title": "EfficientNet Weights for Keras", "url": "/competitions/siim-isic-melanoma-classification/discussion/154322", "author_name": "", "post_date": "2020-05-28T01:42:12.104872300Z", "votes": 3, "comment_count": 1, "views": 0, "content": "
\n

EfficientNets are a family of image classification models, which achieve state-of-the-art accuracy, yet being an order-of-magnitude smaller and faster than previous models. (Source)

\n
\n\n

EfficientNets were used in the best solutions of the latest Kaggle Computer Vision competitions. With the dataset I created thanks to @pavel92 EfficientNet repository, I hope to make easy the development of TensorFlow and Keras models using EfficientNets on Kaggle. Especially I hope to see models using both TF / Keras EfficientNets and TPUs!

\n\n

Here is the link to the dataset. It contains all EfficientNet weights with Baseline, AutoAugment, AdvProp and NoisyStudent preprocessing.

\n\n

Your feedback is highly appreciated if there are ways for me to improve this dataset!

\n\n

Good luck to all and enjoy this new Computer Vision competition! 🎉

", "messages": [ { "id": "864408", "postDate": "05/28/2020 01:42:12", "content": "
\n

EfficientNets are a family of image classification models, which achieve state-of-the-art accuracy, yet being an order-of-magnitude smaller and faster than previous models. (Source)

\n
\n\n

EfficientNets were used in the best solutions of the latest Kaggle Computer Vision competitions. With the dataset I created thanks to @pavel92 EfficientNet repository, I hope to make easy the development of TensorFlow and Keras models using EfficientNets on Kaggle. Especially I hope to see models using both TF / Keras EfficientNets and TPUs!

\n\n

Here is the link to the dataset. It contains all EfficientNet weights with Baseline, AutoAugment, AdvProp and NoisyStudent preprocessing.

\n\n

Your feedback is highly appreciated if there are ways for me to improve this dataset!

\n\n

Good luck to all and enjoy this new Computer Vision competition! 🎉

", "rawMarkdown": "> EfficientNets are a family of image classification models, which achieve state-of-the-art accuracy, yet being an order-of-magnitude smaller and faster than previous models. ([Source](https://github.com/tensorflow/tpu/tree/master/models/official/efficientnet))\n\nEfficientNets were used in the best solutions of the latest Kaggle Computer Vision competitions. With the dataset I created thanks to @pavel92 EfficientNet [repository](https://github.com/qubvel/efficientnet), I hope to make easy the development of TensorFlow and Keras models using EfficientNets on Kaggle. Especially I hope to see models using both TF / Keras EfficientNets and TPUs!\n\n[Here is the link to the dataset](https://www.kaggle.com/mika30/efficientnet-weights-for-keras). It contains all EfficientNet weights with Baseline, AutoAugment, AdvProp and NoisyStudent preprocessing.\n\nYour feedback is highly appreciated if there are ways for me to improve this dataset!\n\nGood luck to all and enjoy this new Computer Vision competition! 🎉", "votes": null }, { "id": "864500", "postDate": "05/28/2020 03:22:03", "content": "

Hi, thanks for making such great datasets. I've also made two similar for noisy students and in general. :)

", "rawMarkdown": "Hi, thanks for making such great datasets. I've also made two similar for [noisy students](https://www.kaggle.com/ipythonx/efficientnet-keras-noisystudent-weights-b0b7) and [in general](https://www.kaggle.com/ipythonx/keras-pretrained-imagenet-weights). :)", "votes": null } ], "comments": [ { "id": 864500, "author_name": "ipythonx", "author_url": "", "post_date": "05/28/2020 03:22:03", "content": "

Hi, thanks for making such great datasets. I've also made two similar for noisy students and in general. :)

", "votes": null, "replies": [] } ], "raw_markdown_by_id": { "864408": "> EfficientNets are a family of image classification models, which achieve state-of-the-art accuracy, yet being an order-of-magnitude smaller and faster than previous models. ([Source](https://github.com/tensorflow/tpu/tree/master/models/official/efficientnet))\n\nEfficientNets were used in the best solutions of the latest Kaggle Computer Vision competitions. With the dataset I created thanks to @pavel92 EfficientNet [repository](https://github.com/qubvel/efficientnet), I hope to make easy the development of TensorFlow and Keras models using EfficientNets on Kaggle. Especially I hope to see models using both TF / Keras EfficientNets and TPUs!\n\n[Here is the link to the dataset](https://www.kaggle.com/mika30/efficientnet-weights-for-keras). It contains all EfficientNet weights with Baseline, AutoAugment, AdvProp and NoisyStudent preprocessing.\n\nYour feedback is highly appreciated if there are ways for me to improve this dataset!\n\nGood luck to all and enjoy this new Computer Vision competition! 🎉", "864500": "Hi, thanks for making such great datasets. I've also made two similar for [noisy students](https://www.kaggle.com/ipythonx/efficientnet-keras-noisystudent-weights-b0b7) and [in general](https://www.kaggle.com/ipythonx/keras-pretrained-imagenet-weights). :)" }, "source": "meta" }