![](https://img-blog.csdnimg.cn/20201014180756927.png?x-oss-process=image/resize,m_fixed,h_64,w_64)
算法论文
文章平均质量分 76
tiaozhanzhe1900
FPGA
展开
-
文献阅读(104)Transformer
题目:An Image is Worth 16x16 Words: Transformers for Image Recognition at Scale时间:2021会议:ICLR研究机构:谷歌原创 2022-05-20 01:15:42 · 67 阅读 · 0 评论 -
文献阅读(83)HAQ
文章目录1 abstract & introduction &related work2 approach2.1 observation(state space)2.2 action space2.3 硬件的直接反馈2.4 量化2.5 reward function奖励函数2.6 agent题目:HAQ: Hardware-Aware Automated Quantization with Mixed Precision时间:2019会议:IEEE Conference on Com原创 2020-03-20 11:03:01 · 383 阅读 · 0 评论 -
文献阅读(100)随机计算神经网络
文章目录1 缩写 & 引用2 motivation & introduction3 实现方法3.1 二进制转概率域1 introduction1 abstract & introduction题目:Neural Synaptic Plasticity-Inspired Computing: AHigh Computing Efficient Deep Convolutional Neural Network Accelerator时间:2020期刊:TCAS1研究机构:成原创 2021-02-20 22:22:16 · 227 阅读 · 0 评论 -
文献阅读(103)SNN
文章目录名词解释题目:Deep learning in spiking neural networks时间:2019期刊:Neural Network(B类)研究机构:路易斯安那大学拉法叶分校名词解释spike-timing-dependent plasticity(STDP)Neuroscientists have identified many variants of this learning rule that falls under the umbrella term sp原创 2021-07-10 11:18:55 · 469 阅读 · 0 评论 -
文献笔记(101)动态神经网络
文章目录1 abstract & introduction2 方法题目:Thanks for Nothing: Predicting Zero-Valued Activations with Lightweight Convolutional Neural Networks时间:2020会议:ECCV研究机构:以色列理工学院1 abstract & introduction根据输出特征图的相邻值,动态预测输出特征图是否为零值,从而减少了卷积运算的次数。Motivation:原创 2021-04-17 15:55:11 · 302 阅读 · 0 评论 -
文献阅读(82)二值化神经网络
文章目录1 abstract & introduction2 Clamping Rectified Linear Unit3 量化方法3.1 前向近似3.2 后向近似题目:Efficient Quantization for Neural Networks with Binary Weights and Low Bitwidth Activations时间:2019会议:AAAI...原创 2020-07-05 08:57:40 · 3052 阅读 · 0 评论 -
文献阅读(90)
文章目录1 abstract & introduction2 方法1 abstract & introduction2 方法3 实验题目:Two-Step Quantization for Low-bit Neural Networks时间:2018会议:CVPR研究机构:中科大/阿里参考博客:https://blog.csdn.net/qq_19784349/article/details/83931420参考博客:https://blog.csdn.net/JachinMa原创 2020-06-10 11:07:02 · 444 阅读 · 0 评论 -
文献阅读(96)目标检测综述(部分)
文章目录1 Detection Components2 人脸识别3 行人识别数据集题目:Recent Advances in Deep Learning for Object Detection时间:2019网络:arXiv研究机构:新加坡管理学院1 Detection ComponentsFPN: Feature Pyramid Networks特征金字塔FPN comb...原创 2020-07-15 17:02:38 · 295 阅读 · 0 评论 -
文献阅读(98)量化
文章目录1 缩写 & 引用2 abstract & introduction题目:Quantization and Training of Neural Networks for Efficient Integer-Arithmetic-Only Inference时间:2018会议:CVPR研究机构:谷歌参考链接1:https://blog.csdn.net/mizh...原创 2020-10-03 23:11:17 · 230 阅读 · 0 评论 -
文献阅读(99)韩松2016年网络压缩论文
文章目录1 缩写 & 引用2 abstract & introduction题目:Deep Compression compressing deep neural networks with pruning, train quantization and huffman coding时间:2016会议:ICLR研究机构:韩松1 缩写 & 引用2 abstra...原创 2019-12-22 18:21:57 · 297 阅读 · 0 评论 -
文献阅读(95)AutoML+剪枝
文章目录1 缩写 & 引用2 abstract & introduction & related work3 方法论题目:AMC: AutoML for Model Compression and Acceleration on Mobile Devices时间:2018会议:ECCV研究机构:MIT韩松1 缩写 & 引用AMC: AutoML f...原创 2019-12-04 13:51:47 · 259 阅读 · 0 评论 -
文献阅读(91)韩松 proxyless NAS
文章目录1 缩写 & 引用2 abstract & introduction3题目:proxyless NAS: direct neural architecture search on target task and hardware时间:2019会议:ICLR研究机构:韩松1 缩写 & 引用2 abstract & introduction强化...原创 2020-02-14 09:22:09 · 249 阅读 · 0 评论 -
文献阅读(102)SAR船舰检测
文章目录1 abstract & introduction2 方法2.1 特征融合2.2 特征增强2.3 特征金字塔1 abstract & introduction2 补充2.1 Shuffle-group enhance (SSE)2.2 CenterNet题目:ShipDeNet-20: An Only 20 Convolution Layers and <1-MB Lightweight SAR Ship Detector期刊:IEEE Geoscience and Re原创 2021-04-28 19:36:34 · 70 阅读 · 0 评论