Timezone: »
Innovations in neural architectures have fostered significant breakthroughs in language modeling and computer vision. Unfortunately, novel architectures often result in challenging hyper-parameter choices and training instability if the network parameters are not properly initialized. A number of architecture-specific initialization schemes have been proposed, but these schemes are not always portable to new architectures. This paper presents GradInit, an automated and architecture agnostic method for initializing neural networks. GradInit is based on a simple heuristic; the norm of each network layer is adjusted so that a single step of SGD or Adam with prescribed hyperparameters results in the smallest possible loss value. This adjustment is done by introducing a scalar multiplier variable in front of each parameter block, and then optimizing these variables using a simple numerical scheme. GradInit accelerates the convergence and test performance of many convolutional architectures, both with or without skip connections, and even without normalization layers. It also improves the stability of the original Transformer architecture for machine translation, enabling training it without learning rate warmup using either Adam or SGD under a wide range of learning rates and momentum coefficients. Code is available at https://github.com/zhuchen03/gradinit.
Author Information
Chen Zhu (University of Maryland, College Park)
Renkun Ni (University of Maryland)
Zheng Xu (Google AI)
Kezhi Kong (University of Maryland, College Park)
W. Ronny Huang (Google Research)
Tom Goldstein (Rice University)
More from the Same Authors
-
2020 : Data Augmentation for Meta-Learning »
Renkun Ni -
2021 : A Closer Look at Distribution Shifts and Out-of-Distribution Generalization on Graphs »
Mucong Ding · Kezhi Kong · Jiuhai Chen · John Kirchenbauer · Micah Goldblum · David P Wipf · Furong Huang · Tom Goldstein -
2021 : Diurnal or Nocturnal? Federated Learning from Periodically Shifting Distributions »
Chen Zhu · Zheng Xu · Mingqing Chen · Jakub Konečný · Andrew S Hard · Tom Goldstein -
2021 : Learning Revenue-Maximizing Auctions With Differentiable Matching »
Michael Curry · Uro Lyi · Tom Goldstein · John P Dickerson -
2021 : Learning Revenue-Maximizing Auctions With Differentiable Matching »
Michael Curry · Uro Lyi · Tom Goldstein · John P Dickerson -
2023 Poster: Battle of the Backbones: A Large-Scale Comparison of Pretrained Models across Computer Vision Tasks »
Micah Goldblum · Hossein Souri · Renkun Ni · Manli Shu · Viraj Prabhu · Gowthami Somepalli · Prithvijit Chattopadhyay · Adrien Bardes · Mark Ibrahim · Judy Hoffman · Rama Chellappa · Andrew Wilson · Tom Goldstein -
2022 Poster: Pre-Train Your Loss: Easy Bayesian Transfer Learning with Informative Priors »
Ravid Shwartz-Ziv · Micah Goldblum · Hossein Souri · Sanyam Kapoor · Chen Zhu · Yann LeCun · Andrew Wilson -
2021 : A Closer Look at Distribution Shifts and Out-of-Distribution Generalization on Graphs »
Mucong Ding · Kezhi Kong · Jiuhai Chen · John Kirchenbauer · Micah Goldblum · David P Wipf · Furong Huang · Tom Goldstein -
2021 Poster: Can You Learn an Algorithm? Generalizing from Easy to Hard Problems with Recurrent Networks »
Avi Schwarzschild · Eitan Borgnia · Arjun Gupta · Furong Huang · Uzi Vishkin · Micah Goldblum · Tom Goldstein -
2021 Poster: VQ-GNN: A Universal Framework to Scale up Graph Neural Networks using Vector Quantization »
Mucong Ding · Kezhi Kong · Jingling Li · Chen Zhu · John Dickerson · Furong Huang · Tom Goldstein -
2021 Poster: Gradient-Free Adversarial Training Against Image Corruption for Learning-based Steering »
Yu Shen · Laura Zheng · Manli Shu · Weizi Li · Tom Goldstein · Ming Lin -
2021 Poster: Adversarial Examples Make Strong Poisons »
Liam Fowl · Micah Goldblum · Ping-yeh Chiang · Jonas Geiping · Wojciech Czaja · Tom Goldstein -
2021 Poster: Center Smoothing: Certified Robustness for Networks with Structured Outputs »
Aounon Kumar · Tom Goldstein -
2021 Poster: Encoding Robustness to Image Style via Adversarial Feature Perturbations »
Manli Shu · Zuxuan Wu · Micah Goldblum · Tom Goldstein -
2021 Poster: Long-Short Transformer: Efficient Transformers for Language and Vision »
Chen Zhu · Wei Ping · Chaowei Xiao · Mohammad Shoeybi · Tom Goldstein · Anima Anandkumar · Bryan Catanzaro -
2020 : W Ronny Huang---Understanding Generalization through Visualizations »
W. Ronny Huang -
2020 : The Intrinsic Dimension of Images and Its Impact on Learning »
Chen Zhu · Micah Goldblum · Ahmed Abdelkader · Tom Goldstein · Phillip Pope -
2020 Poster: MetaPoison: Practical General-purpose Clean-label Data Poisoning »
W. Ronny Huang · Jonas Geiping · Liam Fowl · Gavin Taylor · Tom Goldstein -
2020 Poster: Large-Scale Adversarial Training for Vision-and-Language Representation Learning »
Zhe Gan · Yen-Chun Chen · Linjie Li · Chen Zhu · Yu Cheng · Jingjing Liu -
2020 Spotlight: Large-Scale Adversarial Training for Vision-and-Language Representation Learning »
Zhe Gan · Yen-Chun Chen · Linjie Li · Chen Zhu · Yu Cheng · Jingjing Liu -
2019 : Break / Poster Session 1 »
Antonia Marcu · Yao-Yuan Yang · Pascale Gourdeau · Chen Zhu · Thodoris Lykouris · Jianfeng Chi · Mark Kozdoba · Arjun Nitin Bhagoji · Xiaoxia Wu · Jay Nandy · Michael T Smith · Bingyang Wen · Yuege Xie · Konstantinos Pitas · Suprosanna Shit · Maksym Andriushchenko · Dingli Yu · Gaël Letarte · Misha Khodak · Hussein Mozannar · Chara Podimata · James Foulds · Yizhen Wang · Huishuai Zhang · Ondrej Kuzelka · Alexander Levine · Nan Lu · Zakaria Mhammedi · Paul Viallard · Diana Cai · Lovedeep Gondara · James Lucas · Yasaman Mahdaviyeh · Aristide Baratin · Rishi Bommasani · Alessandro Barp · Andrew Ilyas · Kaiwen Wu · Jens Behrmann · Omar Rivasplata · Amir Nazemi · Aditi Raghunathan · Will Stephenson · Sahil Singla · Akhil Gupta · YooJung Choi · Yannic Kilcher · Clare Lyle · Edoardo Manino · Andrew Bennett · Zhi Xu · Niladri Chatterji · Emre Barut · Flavien Prost · Rodrigo Toro Icarte · Arno Blaas · Chulhee Yun · Sahin Lale · YiDing Jiang · Tharun Kumar Reddy Medini · Ashkan Rezaei · Alexander Meinke · Stephen Mell · Gary Kazantsev · Shivam Garg · Aradhana Sinha · Vishnu Lokhande · Geovani Rizk · Han Zhao · Aditya Kumar Akash · Jikai Hou · Ali Ghodsi · Matthias Hein · Tyler Sypherd · Yichen Yang · Anastasia Pentina · Pierre Gillot · Antoine Ledent · Guy Gur-Ari · Noah MacAulay · Tianzong Zhang -
2019 : Coffee/Poster session 1 »
Shiro Takagi · Khurram Javed · Johanna Sommer · Amr Sharaf · Pierluca D'Oro · Ying Wei · Sivan Doveh · Colin White · Santiago Gonzalez · Cuong Nguyen · Mao Li · Tianhe Yu · Tiago Ramalho · Masahiro Nomura · Ahsan Alvi · Jean-Francois Ton · W. Ronny Huang · Jessica Lee · Sebastian Flennerhag · Michael Zhang · Abram Friesen · Paul Blomstedt · Alina Dubatovka · Sergey Bartunov · Subin Yi · Iaroslav Shcherbatyi · Christian Simon · Zeyuan Shang · David MacLeod · Lu Liu · Liam Fowl · Diego Mesquita · Deirdre Quillen -
2019 Poster: Adversarial training for free! »
Ali Shafahi · Mahyar Najibi · Mohammad Amin Ghiasi · Zheng Xu · John Dickerson · Christoph Studer · Larry Davis · Gavin Taylor · Tom Goldstein -
2018 Poster: Poison Frogs! Targeted Clean-Label Poisoning Attacks on Neural Networks »
Ali Shafahi · W. Ronny Huang · Mahyar Najibi · Octavian Suciu · Christoph Studer · Tudor Dumitras · Tom Goldstein -
2018 Poster: Visualizing the Loss Landscape of Neural Nets »
Hao Li · Zheng Xu · Gavin Taylor · Christoph Studer · Tom Goldstein -
2017 Poster: Training Quantized Nets: A Deeper Understanding »
Hao Li · Soham De · Zheng Xu · Christoph Studer · Hanan Samet · Tom Goldstein