Deploying convolutional neural networks (CNNs) on embedded devices is difficult due to the limited memory and computation resources. The redundancy in feature maps is an important characteristic of those successful CNNs, but has rarely been investigated in neural architecture design. This paper proposes a novel Ghost module to generate more feature maps from cheap operations. Based on a set of intrinsic feature maps, we apply a series of linear transformations with cheap cost to generate many ghost feature maps that could fully reveal information underlying intrinsic features.
2019: Kai Han, Yunhe Wang, Qi Tian, Jianyuan Guo, Chunjing Xu, Chang Xu
https://arxiv.org/pdf/1911.11907v2.pdf
Architecture-Agnostic Masked Image Modeling - From ViT back to CNN
Learning the Beauty in Songs: Neural Singing Voice Beautifier
LiT: Zero-Shot Transfer with Locked-image Text Tuning
Large-Scale Intelligent Microservices
Collaborative Neural Rendering using Anime Character Sheets
Alpa: Automating Inter- and Intra-Operator Parallelism for Distributed Deep Learning
Can Wikipedia Help Offline Reinforcement Learning?
Masked Autoencoders that Listen
Reconstructing 3D Human Pose by Watching Humans in the Mirror
A Conversational Paradigm for Program Synthesis
Masked Siamese Networks for Label-Efficient Learning
Multiface: A Dataset for Neural Face Rendering
OCR-free Document Understanding Transformer
OpenXAI: Towards a Transparent Evaluation of Model Explanations
YOLOv7: Trainable bag-of-freebies sets new state-of-the-art for real-time object detectors
Matryoshka Representations for Adaptive Deployment
Scaling Up Your Kernels to 31x31: Revisiting Large Kernel Design in CNNs
More ConvNets in the 2020s: Scaling up Kernels Beyond 51x51 using Sparsity
The Web Is Your Oyster - Knowledge-Intensive NLP against a Very Large Web Corpus
Join Podbean Ads Marketplace and connect with engaged listeners.
Advertise Today
Create your
podcast in
minutes
It is Free
Babbage from The Economist
The WAN Show
Cybersecurity Today
The 404 Media Podcast
Geek Warning