LiuDongdong

爱好由来落笔难,一诗千改心始安。

Seq2seq_Introduce

A sequence-to-sequence model is a model that takes a sequence of items (words, letters, features of an images…etc) and outputs another sequence of items. The encoder processes each item in the input sequence, it compiles the information it captures into a vector (called the context). After processing the entire input sequence, the encoder sends the context over to the decoder,

Template

level: author: date: keyword: Paper: Summary Research Objective Application Area: Purpose: Proble Statement previous work: Methods Problem Formulation: system overview: 【Qustion 1】 Evaluation Environment: Dataset: Conclusion Notes 去加强了解 Paper: LIRe Summary 现象,辅助支撑材料,目前按解决方案不足, 介于此,我们想

Wavelet

Fourier Transform is the dot product between real signal and various frequency of sine wave. We get a stats of frequency but we don’t know when that “frequency” happen, we lost the time resolution of the real signal. To get both frequency and time resolution we can be dividing the original signal into several parts and apply Fourier Transform to

weekmodel

1. KeypointDetection 1.1. CharPointDetection 识别字符中的俩个关键点。 1.2. Facial-keypoints-detection 用于检测人脸的68个关键点示例。 1.3. Hourglass-facekeypoints 使用基于论文Hourglass 的模型实现人体关键点检测。 1.4. Realtime-Action-Recognition containing: - boxes (Tensor[N,
0%