Bale 1 proposed an end-to-end image compression model based on a variational autoencoder, which adopted a superpriori scheme combining side information. The model is shown...
OpenPPL is SenseTime's open source deep learning reasoning platform based on self-developed high-performance operator library, which enables artificial intelligence applications to run efficiently and reliably...
Abstract: In this paper, we demonstrate that CLAS, an end-to-end contextual ASR model consisting of a full neural network, fuses contextual information by mapping all...
This rendering method will have its own coordinate system, and this coordinate system will occupy the real pixel points of the image, resulting in the...
Guided Anchoring solves the problems existing in the conventional manual preset anchor by generating the anchor online, and provides two implementation methods in terms of...
On September 9, 2021.21, Zhejiang Province held the 7th Engineering Training Competition. Our group participated in the project of garbage classification, and our group successfully...
Recently, Baidu Ernie upgraded to 3.0, heavy release of knowledge enhancement of tens of billions of parameters large model. This model not only learns vocabulary,...
Abstract: In a convolutional neural network, different features are extracted by using filters whose weights are automatically learned during training, and then all these extracted...
In this paper, the performance of CornerNet is optimized, and two optimized varieties of CornerNet are proposed, namely cornernet-saccade and cornernet-squeeze. The optimized means are...
Abstract: This article takes you to experience the application of quantum neural network in natural language processing from zero. This article is shared from Huawei...
The design idea of RepPoints is very clever, using the point set rich in semantic information to represent the target, and skillfully using deformable convolution...
1) Since soft thresholding is a common step of signal denoising algorithm, the deep residual shrinkage network is more suitable for data with strong noise...
The design of Densebox detection algorithm is very advanced, and many anchor-free methods now have their own shadow. If they did not appear a little...
SACCAdenet carries out preliminary target positioning based on the center point feature, and then fine-tune the prediction frame by using the corner point feature and...
The core of CentripetalNet lies in the new corner matching method. An additional centripetal offset value is learned, and a corner with a sufficiently small...
FSAF makes an in-depth analysis of the selection of FPN layer during training, and embedding the original network in the form of ultra-simple anchor-free branching...
Recently, the first Graph Neural Network Competition jointly organized by KDD Cup and OGB (Open Graph Benchmark) was officially released. In the fierce competition between...
Abstract: With the widespread use of administrative documents used to convey and record business information, methods for automating and understanding content from these documents in...
Face detection is the first step of face recognition and processing, mainly used to detect and locate the face in the picture, return high-precision face...
This article is intended to help those who want to gain deep learning practice experience through the TensorFlow Eager pattern. TensorFlow Eager allows you to...
Last year, we launched Transformer, a new machine learning model that outperforms existing machine translation algorithms and other language understanding tasks. Prior to Transformer, most...