Blog | Divakar Verma | Computer Science

MNIST Muddle

Creating poorly written numeric digits using AutoEncoders and MNIST dataset. Demo website - LINK GitHub link - LINK Contents: Project aim...

H.O.M.L Ch-4 | Training models algo

Uncovering some of the black box algorithms in this chapter. Mainly we'll be talking about Linear Regression, Polynomial Regression,...

Azure Cognitive services using Python & REST API | Text-to-Speech

Python script to access Azure's Text-to-Speech cognitive API. Converts all sentences from a file to respective speech audios.

H.O.M.L Chapter-3 | Classification

Turning our attention to Classification systems. Contents: Performance Measure for Classification Precision Recall F1 score...

H.O.M.L Ch-2 notes | End-to-end ML Project

Contents: Root mean square error vs Mean absolute error Few points on test set preparation Correlation Data cleaning caution Text and...

H.O.M.L Chp-1 Summary | The ML Landscape

[1]Types of ML Systems. [2]Main Challenges of Machine Learning. [3]Testing and Validation.

CTC Loss (part-2) | Forward pass using alpha matrix

In this post we will see how CTC Loss is efficiently calculated using Dynamic Programming. We do that by creating a 2D matrix, known as...

CTC Loss for un-segmented Data

Its importance & where it is used: It is used for dealing with un-segmented sequence data. Such data is ubiquitous and it may not be...

Restoring trained weights after modifying the model graph in TensorFlow

Background: Recently, I had a situation where my trained model (TF) could not run on a platform because few nodes in the graph...

Batch Normalization - 1 | Using tf.nn.batch_normalization

See how to correctly use batch normalization in TensorFlow using tf.nn.batch_normalization. User needs to properly update the values duri...

Post init

This is my first blog post. Ready to share the world what I learn everyday. Show love if you enjoy the posts. -Divakar