Facing Value Error ,while trying to fit sigmoid nueron in ,DL #104 Feed Forward Neural Network

Curriculum Help Deep Learning
#1

I am facing the value error : Input contains NaN, infinity or a value too large for dtype(‘float64’).

while I was going through the DL#104 subsection of ML course at GUVI.in.I have provided the link for my colab notebook.Please help me sort out my mistake.https://colab.research.google.com/drive/12J47nVxMT3kcqt24VsPVfIC8LfITvHSn?usp=sharing

#2

Its because of sigmoid function:

def sigmoid(self,x):
    return 1.0/1.0 - np.exp(-x)

Verify the brackets and signs.

#3

I verified my sigmoid neuron function with yours and even replaced it as well,but am still facing the same error.

#4

Hi @divyansh_arora,
Please replace this snippet with the following:

return 1.0/(1.0 - np.exp(-x))
#5

I tried doing the change but am facing a Type Error along with a Runtime warning.

#6

Hi,
For overflow warning, please refer the following thread: Overflow encountered in exp.
For TypeError, Please refer TypeError: Dict object is not callable.