ValueError: Input contains NaN, infinity or a value too large

[Anand_Deep]

hello @everyone
When I am trying to fit the model in Feed Forward Neural Network(Lecture name: “DNN using python: Classification with sigmoid neuron” on Guvi) it raises an error which says :

ValueError: Input contains NaN, infinity or a value too large for dtype('float64').

[Ishvinder]

Hi @Anand_Deep, Please do share a link to your notebook with “Anyone can access” permission.

[Anand_Deep]

notebook link
please resolve

[Ishvinder]

Hi @Anand_Deep,
Please normalize the data before feeding it.

[Anand_Deep]

now when i am plotting the curve it’s giving me straight line.
Notebook Link

[sanjayk]

@Anand_Deep

now when i am plotting the curve it’s giving me straight line.

Inside sigmoid function change

    return 1.0/1.0 + (np.exp(-x))

to

    return 1.0/(1.0 + np.exp(-x))