Optimization_methods : the accuracy evaluated by adam’s algorithm doesn’t match the one reported in later table.
Hi @Ucar97 !
Is it possible you didn’t follow the data cleaning or feature engineering process in the notebook? These steps will affect the performance as well.
mmm… in the next part however the results of the algorithms coincide with those reported in a subsequent table, similar to this one I posted, and I didn’t change anything, I simply proceeded to do the tasks. I think there is an error in this table I posted
Here’s what it looks like in my notebook:
So my conclusion would be that there is something wrong with your Adam code. I got 94.3% accuracy which is consistent with the table.
2 Likes
Thank you, I have to find this mistake
1 Like
Please let us know how it goes.
Hi Paul, I got 94.33 for Gradient descent, 95.33 for Momentum, 94% for Adam, but it shows in the notebook that
it seems the first two methods should be > 94.6 and > 95.6. Can I ask what might be the issue here?
The numbers I get are the same as yours. I don’t think this is a problem, but maybe we should get them to adjust the published expectations here.
1 Like