Can I use this graph as my final result?

1

I have got this graph while training my NN, but I am not getting whether I am doing right or not? Shall I consider the final result or this dip will create problem in analysis. Please guide.

Figure 2021-09-18 002504

2

Depends on what the analysis is for, i would consider the plot to be sufficient. As for the dip could be cause of difference reasons:

  1. Something returning NaN
  2. Randomness (unlikely but possible)
3

Thanks ! What do you mean by NaN, I mean where it is returning NaN values?
Can I show this graph in my paper and say My accuracy is 94% which is my final result of classification, ignoring this dip.

4

I have never encountered such an error and i’m not the best person this :sweat_smile: I would probably check with a supervisor about including this in a paper and the best supportive argument

5

Hi @arya47 did you ask the supervisor?