What's new in Tensorflow 2?

-
I will highlight some of the key changes with the latest version of Tensorflow 2.0 compared to earlier versions:
  • Now Keras users who use multi-backend Keras with the TensorFlow backend are needed switch to tf.keras in TensorFlow 2.0. As it’s going to be better maintained.
  • By default Tensorflow 2.0 will be having Eager execution that evaluates operations immediately, without building graphs. You can obviously switch to the style of building graphs, as it was in Tensorflow earlier as well
  • It is now having a full lower-level API, which allows it to use for having custom training loops(which is needed by researchers a lot)
  • With tf.GradientTape you can now do automatic differentiation
  • One of the other features of Tensorflow is experimental support for mixed precision on GPUs and Cloud TPUs
  • A clear consistent, intuitive, syntax across various APIs
  • Tensorflow 2.1 bought support for TPU training in a simple and intuitive way which makes it’s a favourite for Kagglers like me
  • Internals operations are now accessible by tf.raw_ops
  • Inheritable interfaces for variables, checkpoints, layers

Comments

Post a Comment

Popular posts from this blog

Scenario based interview questions for Data science!

-
Image
  Q1. You are given a train data set having 1000 columns and 1 million rows. The data set is based on a classification problem. Your manager has asked you to reduce the dimension of this data so that model computation time can be reduced. Your machine has memory constraints. What would you do? (You are free to make practical assumptions.) Answer: Processing a high dimensional data on a limited memory machine is a strenuous task, your interviewer would be fully aware of that. Following are the methods you can use to tackle such situation: Since we have lower RAM, we should close all other applications in our machine, including the web browser, so that most of the memory can be put to use. We can randomly sample the data set. This means, we can create a smaller data set, let’s say, having 1000 variables and 300000 rows and do the computations. To reduce dimensionality, we can separate the numerical and categorical variables and remove the correlated variables. For numerical variable...
Read more

What is Data Preprocessing?

-
Image
  Data preprocessing  is a step in the data mining and data analysis process that takes raw data and transforms it into a format that can be understood and analyzed by computers and machine learning. When we talk about  data , we usually think of some  large datasets  with some rows and columns. While that is a likely scenario, it is not always the case — data could be in so many different forms:  Structured Tables, Images, Audio files, Videos , etc. Machines don’t understand text, image, or video data as it is, they understand 1s and 0s. Real-world data also contains noises, missing values, etc. which cannot be directly used for ML models. Hence,  data preprocessing  is required for cleaning the data and making it suitable for an ML model which increases the accuracy and efficiency of the model. It involves the following steps: Getting the Dataset Importing Libraries Importing Dataset Data Quality Assessment: i) Finding and Processing Missing/Inc...
Read more