• Huawei’s latest overview of visual transformer with Peking University and Sydney University


    Author Yu JunReport | I love computer vision (wechat ID: aicvml) Transformer technology first originated in the field of natural language processing, but in May this year, a Facebook article applied it to target detection in computer vision (Detr algorithm, which has been cited 78 times) to make it shine, and quickly attracted the attention […]

  • Netty sticky package & half package solution


    Short link Disconnect after sending data and re-establish the connection when sending next time //Client side Fixed length decoder When the length of the data sent by the client is fixed, set the decoder of the server as the fixed length decoderFixedLengthFrameDecoderThe problem of sticking package and half package can be avoided //Server side Row […]

  • Using depth convolution self encoder to reduce image noise in 10 minutes


    Author | Orhan Gazi YAL ç ı nvCompile VKSource: towards Data Science You may be familiar with different neural network structures. You may have heard of feedforward neural networks, CNN, RNNs. These neural networks are very useful for solving supervised learning tasks such as regression and classification. However, in the field of unsupervised learning, we […]

  • Current situation and future of webrtc (Part 2)


    Author / Chad Hart Original link/https://webrtchacks.com/webrt… Scalable video coding Scalable video coding (SVC) can be said to be a better way to process multiple media streams from the same sender to deal with different conditions of each receiver in a group call. In many ways, it is also considered more complex. Sergio & Gustavo published […]

  • The application of insight style attention network (sanet) in arbitrary style transfer


    By dhwani MehtaCompile FlinSource: Media Image stylization is a kind of image processing technology studied in recent decades. This paper aims to show an efficient and novel style attention network (sanet) method, which can balance the global and local style patterns while preserving the content structure and synthesizing high-quality stylized images. An overview of style […]

  • MIME type encyclopedia, you are worth collecting


    [please indicate the source for reprint]:https://blog.csdn.net/huahao1989/article/details/107746738 Multipurpose Internet Mail Extension (MIME) typesIs a standardized way to represent the nature and format of a document. It’s inIETF RFC 6838It is defined and standardized in. Internet number distribution organization(IANA)Is the official organization responsible for tracking all official MIME typesmedia typePage to find the latest complete list. Browsers […]

  • Intuitive interpretation of neural machine translation


    By ReNu Khandelwal What is neural machine translation? Neural machine translation is a technology that translates one language into another. One example is the conversion of English to Hindi. Let’s think about it. If you’re in an Indian village, most of the people there don’t know English. You plan to communicate with the villagers effortlessly. […]

  • Detr of Facebook AI, a target detection method based on transformer


    By prateek JoshiCompile | VKSource | analytics vidhya introduce Machine learning frameworks or libraries sometimes change the landscape of the field. Today, Facebook has open source such a framework, Detr (detection transformer) In this paper, we will quickly understand the concept of target detection, and then directly study Detr and its benefits. object detection In […]

  • Spring cloud – Chapter 9 feign


    1: What is feign Feign is a declarative web service client, which makes it very easy to write a web service client. You just need to create an interface and add annotations on it. Official website: https://github.com/OpenFeign/… 2: What can feign do Feign is designed to make it easier to write java HTTP clients. When […]

  • Keras implementation of VAE variational auto encoder


    Variational auto encoder (VAE) is a kind of generation model. The training model consists of encoder and decoder. The encoder maps the input samples to a certain low dimensional distribution, which is usually a multivariate Gaussian distribution with independent dimensions. Therefore, the output of the encoder is the mean and logarithmic variance of the Gaussian […]

  • Transformer in NLP


    By ReNu KhandelwalCompile | VKSource: toward Data Science In this article, we will discuss the following questions about transformer Why do we need a transformer, and what are the challenges of the sequence 2 sequence model? Transformer and its architecture are introduced in detail In depth study of the terms used in transformer, such as […]

  • Paper reading: eNet


    introduction Pixel level classification tasks such as semantic segmentation have been developed a lot, and the existing models are developing towards higher and higher accuracy. However, such tasks also have important applications for embedded devices, so real-time semantic segmentation is essential. However, most of the models are towards high-precision, which is very poor in real-time. […]