TensorFlow is the most popular deep learning framework and with the release of Tensorflow 2.0, Keras has been integrated into the TensorFlow ecosystem. This has greatly boosted the ease of use for Tensorflow. Tensorflow provides many amazing functionalities and libraries for deep learning. Tensorboard is one of the amazing features provided by Tensorflow. As per the tensorflow.org:
Tensorboard is an amazing tool for analyzing, visualizing, debugging of training. Tensorboard is part of Tensorflow but can also be installed separately.
The outline of our journey with Tensorflow is as following:
You must have seen someone’s GitHub code and tried to copy that code to try.
We often use Github to showcase the work done to open source community so that others can take benefit. Also, we often fork or pull code from other’s repo for our use. Then there are times when we just want to see how the code will work and fiddle around the code a bit to see if it can fit our requirements, but we end up trying to run notebook in our machines. Binder comes to the rescue.
Tensorflow, the most successful deep learning library is getting a major upgrade with TF 2.0. Tensorflow is changing many of the things with the new release like eager execution will be enabled by default, tf.functions, tf.keras etc. But today we will discuss only about the upgrades in Tensorboard.
Tensorboard is a visualization tool for Tensorflow. With Tensorboard you can monitor the training and other metrics in visual form which helps you get a clear picture of how thing are going on. We will discus the new feature of Tensorboard i.e. Hyperparameter tuning using HParams dashboard.
When you use deep learning…
Tensorflow is a great library for deep learning and has a lot of functionality to offer. Tensorflow provides many features which makes it a lot easier to develop Deep Neural Netwoks.
In this post we are going to see some useful predefined and user defined methods which might help you while working with tensorflow.
This function takes a tensor as an input and returns a tensor with zeros as value and same shape and type as input tensor.
tensor = tf.constant([[1, 2, 3], [4, 5, 6]]) tf.zeros_like(tensor) # [[0, 0, 0], [0, 0, 0]] This function can be helpful in…
Neural networks, the non-living magical computer programs promising a bright future. By Me
AI is the new electricity. By Andrew Ng
You can find tons of great material for learning neural networks. Here we will keep things very simple and will scratch the surface without getting shock ;). So We will see 3 things about neural network:-
This Post is my first post for #100DaysOfML challenge.
Here we assume that you don’t know anything about machine learning or neural networks.
Neural networks are magical programs which try to learn…
When working with deep learning some of the major stoppers are training data and computation resources. Google provides free solution for computation resources with Colaboratory, Google’s free environment which have GPU. Colaboratory provides online jupyter notebook where you can select GPU as back-end server. You can read more about the Colab here.
Here I will add some of the features which might be helpful for you in some scenarios.
Usually to use any file in Colab you need to upload it on Google drive and then use it. …
Deep learning can solve many interesting problems that seems impossible for human, but this comes with a cost, we need a lot of data and computation power to train deep neural networks.
To solve the data problem we can use data Augmentation. Augmentation can also take a lot of computation as we may need to augment millions of images, to handle this we can use tensorflow.image which provides some of the basic image functions and all the computation is done on GPU.
Google just launched the latest version of Tensorflow i.e. Tensorflow 1.7 in Tensorflow Dev Summit 2018. You can check the list of all changes here. Also to watch the full dev summit please visit here. Below are some of the main highlights of TF 1.7:
One of the most interesting feature is eager_execution…
In November Infosys announced that they will train first ever batch of Autonomous Vehicle Engineers with Udacity. Infosys set a target to train around 500 people in 2018 alone. This is a brief dive into the details of the program.
I am proud to be the part of first ever Self Driving Car Nanodegree batch from Infosys. This is equally important program for Infosys and Udacity as this is the first time they are providing in class training for their most prestigious program. Let me take you to the tour of my journey into the program so far.
Autonomous Vehicle Engineer. #FutureAI #LearningAI