Let's compare the execution time of the following function. It adds 1 to a variable 1 billion times. With Numba:

from numba import jit
import time

@jit(nopython=True, cache=True, fastmath={'fast'})
def numba_test():
	x = 0
	for i in range(1000000000):
		x += 1

start = time.time()
numba_test()
print(time.time() - start)

This returned 0.031908512115478516 seconds.

Without Numba:

import time

def test():
	x = 0
	for i in range(1000000000):
		x += 1

start = time.time()
test()
print(time.time() - start)

This returns 23.248998165130615 seconds. As we can see this example of using Numba is 728x times faster.