(the concise version of the invincible version) deeply understand the transformer source code

Original link: https://blog.csdn.net/zhaojc1995/article/details/109276945 reference material: transformer original paper In depth understanding of transformer and source code Graphic Transformer (full version) The Annotated Transformer Chinese annotated version of The Annotated Transformer prefUTF-8...

Posted by Teddy B. on Wed, 29 Sep 2021 08:40:54 +0530

3, Preprocessing of OpenCV image -- binarization and adaptive threshold

Definition: image binarization is to set the gray value of pixels on the image to 0 or 255, that is, to present the whole image with obvious visual effects of only black and white. An image includes target object, background and noise. In order to directly extract the target object from a multiUTF-8...

Posted by bradlybrown on Thu, 30 Sep 2021 02:44:22 +0530

[python learning] python list type

Chinese history [battle of Sakai Quan]: In the process of conquering all ethnic groups in the Central Plains, the Yellow Emperor fought a war with the two tribes of the Yan Emperor in Hanquan. Through this war, the Yellow Emperor defeated the Yan Emperor and replaced the Yan Emperor as the leadUTF-8...

Posted by zipdisk on Thu, 30 Sep 2021 09:31:09 +0530

TypeError: img should be PIL Image. Got <class 'dict'>/Got<class 'Torch.Tensor'>/Got<class 'numpy.n>

1. Problem description When the code runs to the following statement: for i, data in enumerate(train_loader): There are three types of error reports: TypeError: img should be PIL Image. Got TypeError: img should be PIL Image. Got TypeError: img should be PIUTF-8...

Posted by ragy on Fri, 01 Oct 2021 04:02:50 +0530

Review of classic papers in the field of character recognition phase III: FAN

1. Opening With the development of character recognition based on deep learning, it can be classified by decoding methods. Generally, it can be divided into CTC decoding methods and attention decoding methods. Of course, there will be a variety of subdivisions in the specific implementation. ThUTF-8...

Posted by essjay_d12 on Fri, 01 Oct 2021 06:22:44 +0530

Thoughts on quantitative perception training of deep learning framework and a solution of OneFlow

[introduction to GiantPandaCV] this article shares a recent project that the author did in ONEFLOW. After porting Python FX to ONEFLOW, he realized the automatic quantitative perception training dynamic graph model (called NN. Module in both Python and ONEFLOW). Now users can complete the wholeUTF-8...

Posted by ThatGuyBob on Fri, 01 Oct 2021 23:35:29 +0530

[image rain removal] dcsfn: deep cross scale fusion network for single image rainremoval

Code address: GitHub - Ohraincu/DCSFN: DCSFN: Deep Cross-scale Fusion Network for Single Image Rain RemovalDCSFN: Deep Cross-scale Fusion Network for Single Image Rain Removal - GitHub - Ohraincu/DCSFN: DCSFN: Deep Cross-scale Fusion Network for Single Image Rain Removalhttps://github.com/OhraiUTF-8...

Posted by Ryanmcgrim on Sat, 02 Oct 2021 05:49:22 +0530

[Machine Learning] Explain Normalization

Catalog 1. Preface 1.1 Introduction 1.2 Internal Covariate Shift 2. Batch Normalization (BN) Principle of 2.1 BN Details of 2.2 BN Role of 2.2.1 BN How many BN parameters do 2.2.2 convolution layers have? Can 2.2.3 BN have no Scale & Shift process? 2.2.4 BN Layer before or after ReLU? Why is tUTF-8...

Posted by aviatorisu on Sat, 02 Oct 2021 21:50:26 +0530

Pytoch implements the linear model

1. Model training steps (1) Prepare dataset (2) Design model, calculation y_pred (3) Calculate loss (4) Training cycle ① feedforward ② feedback (gradient) ③ update weight 2. Knowledge points (1) Linear Unit ① The linear element is to calculate y once_ PRED, that is, a perceptron. ② In the actuaUTF-8...

Posted by knighthawk1337 on Sun, 03 Oct 2021 00:59:35 +0530

python calls Baidu AI to recognize characters and tables

python calls Baidu AI to recognize characters and tables Get account information 1. You need to register Baidu AI to obtain ID and key. Detailed instructions on registration in Baidu Ai Community: https://ai.baidu.com/forum/topic/show/867951 Just go to "1.6 get key". Then record your APP_ID,APIUTF-8...

Posted by e-novative on Sun, 03 Oct 2021 02:27:58 +0530