# TensorFlow

* An end-to-end open-source platform.
    
    complete solution, from start to finish (end to end), the source code of the platform is freely available (open source).
    
    Tensors:- multi-dimensional array
    
* Based on the concept of computational graphs.
    
    Computational graphs:- a directed graph where nodes -> operations, edges -> data (tensors) flowing between those operations.
    
* Supports eager execution (default mode) and graph execution.
    
    Eager execution:- code is learned immediately after writing it, line by line. Easy debugging and natural control flow.
    
    Graph execution:- after writing the code, build a graph (consisting of a bunch of operations on a bunch of tensors), then learns that graph. Faster execution compared to eager execution.
    

```python
# installing tensorflow library
pip install tensorflow

# importing tensorflow
import tensorflow as tf

# Data storing in tensorflow
# constants
a = tf.constant(5)
print(a)
b = tf.constant(10)
c = tf.add(a, b)
print(c)

# variables
var1 = tf.Variable(20)
print(var1)

# 2D array
var2 = tf.Variable([[1, 2], [3, 4]])
var3 = tf.Variable([[5, 6], [7, 8]])
print(tf.matmul(var2, var3))
```