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
Visual ChatGPT: Talking, Drawing and Editing with Visual Foundation Models
LLaMA: Open and Efficient Foundation Language Models
Dropout Reduces Underfitting
Cross-domain Compositing with Pretrained Diffusion Models
REaLTabFormer: Generating Realistic Relational and Tabular Data using Transformers
Large-scale Multi-Modal Pre-trained Models: A Comprehensive Survey
Fine-Tuning Language Models from Human Preferences
AltCLIP: Altering the Language Encoder in CLIP for Extended Language Capabilities
Demonstrate-Search-Predict: Composing retrieval and language models for knowledge-intensive NLP
Mastering Diverse Domains through World Models
Adding Conditional Control to Text-to-Image Diffusion Models
A Vector Quantized Approach for Text to Speech Synthesis on Real-World Spontaneous Speech
Token Merging: Your ViT But Faster
BioGPT: Generative Pre-trained Transformer for Biomedical Text Generation and Mining
Dual PatchNorm
Reversible Vision Transformers
Offsite-Tuning: Transfer Learning without Full Model
A Length-Extrapolatable Transformer
NaturalSpeech: End-to-End Text to Speech Synthesis with Human-Level Quality
Multimodal Chain-of-Thought Reasoning in Language Models
Join Podbean Ads Marketplace and connect with engaged listeners.
Advertise Today
Create your
podcast in
minutes
It is Free
Cyber Security Headlines
The WAN Show
Software Engineering Daily
Babbage from The Economist
Cybersecurity Today