TensorFlow is open-source software and free data flow and differentiable programming library across a variety of assignments. It is a symbolic math library and is also used for apps for machine learning, such as neural networks. TensorFlow provides desktop, mobile, internet, and cloud-based APIs for beginners and professionals to create.
USE OF IT:
By generating a computational graph, the TensorFlow library enables users to conduct tasks. TensorFlow has taken the world by storm because it is free, (comparatively) simple to use and provides developers with entry-level machine learning backgrounds access to a strong library rather than building all of their AI models from scratch.
WHY IT IS CALLED TensorFlow
Because as a multi-dimensional array, also known as tensors, it requires input. You can build a kind of operations flowchart (called a graph) you want to conduct on that input. The input goes in at one end, and then it flows through this multi-operation scheme and comes out as output from the other end.
It works in three parts:
1. Preprocessing the data
2. Build the model
3. Train and estimate the model
WHERE will RUN?
Categorize hardware and software demands into-
– Development Phase: This is when the mode is being trained. Training generally takes place on your desktop or laptop.
– Run Phase or Inference Phase: Tensorflow can be executed on many separate platforms once training is completed.
a. Desktop running Windows, macOS or Linux
b. Cloud as a web service
c. Mobile devices like iOS and Android
HOW WE OPERATE TensorFlow:
It operates, A graph with distinct computations defines each session. Multiplying by amount can be an easy instance.
Tensorflow requires three steps:
1. Define the variable
2. Define the computation
3. Execute the operation
ALGORITHM LIST :
Currently integrated API for:
a. Linear regression
b. Classification/Deep learning classification
c. Deep learning wipe and profound
d. Booster tree regression and classification
DIFFERENCE BETWEEN KERAS, TensorFlow and PyTorch?
Keras is a Python-based open-source neural network library. It can run at the top of TensorFlow. Intended to allow profound neural networks to experimented quickly.
An open-source software library for programming data flows across a variety of functions. It is a symbolic math library that is used for apps such as neural networks to learn machines.
PyTorch is a fully Python-based open-source machine learning library, can be used for applications like NLP (natural language processing) and was developed by AI (Artificial Intelligence) group of Facebook.