Generalised FFNN, TypeError: Cannot interpret '2' as a data type

self.B[i+1] = np.zeros(1, self.sizes[i+1])

2nd argument tonp.zeros can only be datatype (float, int etc), but here its some number (self.sizes[i+1])

If you share the colab link, it will be easier to pinpoint the reason for error.

its the same file from the previous post

can you have a look at this plot too?

i resolved the previous problem

can you a look at plot its not being classified like it should in the video

Can you put a screenshot here how it looks in the video?

  1. Graph shown in the video and 2) you got : both are far away from being anyway closer to the right prediction. And hence the later sections cover more advanced technique than Sigmoid Neuron.

The difference in your graph and video could be because of: different random initialization or difference in epoch etc.

I will try to look at the code later in the day though I think your result is ok as well considering what a sigmoid neuron can do (trying to separate linearly).

can you have a look at this error message

Can you post this error as a separate new thread and also share the colab link?