首页
技术日记
编程
旅游
数码
登录
标签
Efficient
Transformers are SSMs: Generalized Models and Efficient Algorithms Through Structured State Space Du
这篇论文的标题是《Transformers are SSMs: Generalized Models and Efficient Algorithms Through Structured State Space Duality》。其主要探
Models
Efficient
Generalized
Transformers
SSMS
admin
3月前
37
0
CVPR2022学习-人脸识别:An Efficient Training Approach for Very Large Scale Face Recognition
论文地址: https:arxivpdf2105.10375.pdf 代码地址: GitHub - tiandunxFFC: Official code for fast face classification 看标题大概的理解-
Training
Approach
Efficient
face
Recognition
admin
4月前
39
0
论文阅读:HybridAlpha: An Efficient Approach for Privacy-Preserving Federated Learning
论文名字HybridAlpha: An Efficient Approach for Privacy-Preserving Federated Learning来源会议 the 12th ACM Workshop年份20
论文
Efficient
HybridAlpha
Approach
learning
admin
4月前
56
0
论文阅读 [CVPR-2022] An Efficient Training Approach for Very Large Scale Face Recognition
论文阅读 [CVPR-2022] An Efficient Training Approach for Very Large Scale Face Recognition 一种高效的超大规模人脸识别训练方法 studyai 搜索论文:
论文
Efficient
Training
CVPR
Recognition
admin
4月前
49
0
【Paper Reading】Communication-Efficient Distributed Deep Learning A Comprehensive Survey
Communication-Efficient Distributed Deep Learning: A Comprehensive Survey 原文来源:[Arxiv2020]Communication-Effi
Communication
Efficient
Paper
reading
Distributed
admin
6月前
93
0
Parameter-Efficient Fine-Tuning for Large Models: A Comprehensive Survey
Parameter-Efficient Fine-Tuning for Large Models: A Comprehensive Survey PDF: https:arxivpdf2403.14608.pdf 1 概述 大型
Fine
tuning
parameter
Efficient
Comprehensive
admin
6月前
100
0
Parameter-Efficient Fine-Tuning for Large Models: A Comprehensive Survey阅读笔记
Parameter-Efficient Fine-Tuning for Large Models: A Comprehensive Survey综述阅读笔记仅记录个人比较感兴趣的部分基本知识PEFT的三种分类:a
笔记
Fine
tuning
parameter
Efficient
admin
6月前
105
0
学习笔记 Comprehensive and Delicate: An Efficient Transformer for Image Restoration(CVPR2023)
代码和参考: https:www.zhihuquestion339499743answer3207458947?utm_campaignshareopn&utm_contentgroup1_Ans
学习笔记
Delicate
Comprehensive
Efficient
Restoration
admin
6月前
114
0
LLMs之PEFT:《Parameter-Efficient Fine-Tuning for Large Models: A Comprehensive Survey》翻译与解读
LLMs之PEFT:《Parameter-Efficient Fine-Tuning for Large Models: A Comprehensive Survey》翻译与解读 导读:这篇论文是
Efficient
Fine
parameter
LLMs
PEFT
admin
6月前
122
0
Training data-efficient image transformers & distillation through attention
本视觉Transformers(86M参数)在ImageNet上达到83.1%的top-1精度,蒸馏版本高达84.4%!优于ViT、RegNet和ResNet等,代码刚刚开源! 注:文末附【Transformer】学习交流群Train
Efficient
Image
Training
DATA
distillation
admin
7月前
123
0
论文复现:Learning Efficient Convolutional Networks through Network Slimming
论文核心 论文提出了一种结构化剪枝策略,剪枝对象为 channel ,对 channel 重要性的评价标准使用的是 Batch Normalization 层中的缩放因子,这不会给网络带来额外的开销。 论文细节品读 带 L 1 L1 L
论文
Efficient
learning
Convolutional
Network
admin
7月前
123
0
【论文阅读】DeiT | Training data-efficient image transformers & distillation through attention
本文主要对Facebook最近提出来的DeiT模型进行阅读分析。一、动机:DeiT解决什么问题? 现有的基于Transformer的分类模型ViT需要在海量数据上(JF
论文
Training
DATA
Efficient
DeiT
admin
7月前
125
0
DeiT:Training data-efficient image transformers & distillation through attention
这篇文章主要是通过一些训练策略和知识蒸馏来提升模型的训练速度和性能效果。 原文链接:Training data-efficient image transformers & distillation thro
DATA
Efficient
DeiT
Training
Image
admin
7月前
101
0
Efficient Inference in Fully Connected CRFs with Gaussian Edge Potential
这里我们把每列当成像素, 每行当成不同的label, 这里有四种label. 然后我们需要算在每个点比如第一列第二行的点则为 Q 1 ( x 1第 二 种 l a b e l ) Q_1(x_1第二种label) Q1(x1第二
Fully
Connected
Efficient
Inference
Edge
admin
2025-1-31
86
0