Our dataset is converted into Categorical and hence the softmax activation function will perform better along with categorical loss. Using ADAM has many advantages and one of them is learning rate decay.
DEVANAGARI WRITING DOWNLOAD
Please download the datset from above link and extract them under your current working directory.Ĭonv2d_5 (Conv2D) (None, 28, 28, 32) 9248 Huge credit goes to the team who collected the dataset and made it public. The required dataset is publicly available on the link. I didn’t knew that much about ‘Image Datagenerator’ of Keras then so I converted all the image files to CSV file with first column as label and remaining 1024 as pixel values. The format of image was Grayscale with 2 pixels margin on each side. We could create own version of dataset but why to take lot time rather than working with already collected dataset. The classifier model I used initially was trained on Google Colab, but now for writing of this blog, I am doing it on my own laptop because I am lacking internet access. I would like to give huge credits to the authors of dataset and of course my friends with whom I defended this project. This project was done by me on early 2019 but I had a plan of doing it on 2018 then to complete our academic project of BSc.CSIT, I did it. However, throughout this project I considered them as single character. Some consonants are complex and made by combine some other. I will add a Web App.ĭevanagari is the national font of Nepal and is used widely throughout the India also. Then I will add a new feature to this project.
DEVANAGARI WRITING CODE
Those codes were totally messy and I am thinking of improving the code and entire system. Looking back to those algorithms I applied on that project makes me still fell proud about it but if I looked over the codes, then I have to step back. This has been more than a year I did this project and I have already written a blog about it on medium, and my personal blogs also.
0 kommentar(er)
