Scalar backpropagation multiclass


I created FFSN_MultiClass_Specific class and called fit function the same way as it’s been discussed in video, but in my case ylabels are showing very high values(attached the screenshot). I’m not sure where the mistake is happening. Can you please help me here?

I debug the loss values and it shown like below:

dict_values([1.4722942990161851, 1.4591587671794841, 1.4530155885611429, 1.4479956542180528, 1.4433035655671014, 1.4387889088173924, 1.434401852700895, 1.4301227259804252, 1.4259573394265925, 1.4219358046746604, 1.4181048673417433, 1.4145141801193941, 1.4112016865733292, 1.4081848259448664, 1.4054604462217457, 1.4030106591740774, 1.4008099296174183, 1.3988305064055944, 1.3970456087313905, 1.3954308995317646, 1.3939649381])

Hi @shweta_s,
It seems difficult to figure out the loophole, can you please share your notebook?

I wasn’t able to attach the file here so, I’m sharing the google drive link here. Please check the last class in that notebook i.e. class FFNN_Multiclass. Not sure why ylabels are showing such high values.