Searching for a_robust_neural_architecture_in_four_gpu_hours

Searching for a_robust_neural_architecture_in_four_gpu_hours

1Copyright © Acroquest Technology Co., Ltd. All rights reserved. Searching for A Robust Neural Architecture in Four GPU Hours 2019/6/30 Acroquest Technology Co…

5 mentions: @tereka114@udoooom@jaguring1@kantocv@erukiti
Keywords: gpu
Date: 2019/06/30 05:16

Referring Tweets

@jaguring1 マジ? 「提案手法(GDAS)は 従来の3000GPU時間 → 4GPU時間(0.1%) でNASの実行が完了する。また、従来の手法より精度が高い(複数のSoTA)」 https://t.co/qvVL26EqeE
@tereka114 資料公開しました 「Searching for A Robust Neural Architecture in Four GPU Hours」 https://t.co/SMYIT6mXjg #cvsaisentan
@erukiti すごい。深層学習のネットワークを自動生成するヤツをめっちゃ高速化(ご家庭にある程度のGPUで一日かからない。しかもまだ最適化の余地があるっぽい) / “Searching for a_robust_neural_architecture_in_four_gpu_hours” https://t.co/bgxG2Bcp6Y

Bookmark Comments

Related Entries

Read more Machine Learning Benchmark Set with IBM POWER9 and GPUs
0 users, 0 mentions 2018/04/23 11:41
Read more Mastering Fast Gradient Boosting on Google Colaboratory with free GPU
0 users, 3 mentions 2019/02/27 20:20
Read more MegatronLM: Training Billion+ Parameter Language Models Using GPU Model Parallelism - NVIDIA ADLR
1 users, 23 mentions 2019/08/13 00:00
Read more Misreading ChatEpisode 40Episode 39 – Service Fabric: A Distributed Platform for Building Microservi...
0 users, 0 mentions 2018/06/14 16:30