montana:
  python version: 3.7.3
  keras version: 2.2.4
  Neural network to solve a 1d regression problem.
  The data is read from internal functions.
(100, 2)
(10, 2)
[0.598499 0.60241  0.528659] [-0.40407  -0.542388  1.06106 ]
[2.03206 2.44061 2.37146] [1.3084  1.27157 2.95062]

Training:

step 0: training cost = 4.13479
step 10: training cost = 2.44578
step 20: training cost = 1.50109
step 30: training cost = 0.967201
step 40: training cost = 0.661138
step 50: training cost = 0.482286
step 60: training cost = 0.375155
step 70: training cost = 0.309008
step 80: training cost = 0.266707
step 90: training cost = 0.238612
step 100: training cost = 0.219232

Testing:


10/10 [==============================] - 0s 2ms/step
Test cost: 0.3217616677284241
Weights =  [[1.695848]]
Bias =  [1.4571118]

montana:
  Normal end of execution.