Unable to counter the error

#1

I am trying to make prediction for my test images , but facing, error, and i am not able to counter the error, plz if u can help. Moreover i am also getting an error when i tried to make prediction and check accuracy using inbuilt svm classifier using custom_kernel. Please if u could check the errors in my code?
Code link :- https://drive.google.com/open?id=1w4Bzqa_dS-NTYUcv16RwgGxw8wLs3wB8

#2

Hey @saksham_thukral, when you load testing examples you need to reshape it, into (-1,7500) and than make the prediction.

Your inbuilt svm doesn’t work, because inbuilt svm requires a list of labels, pass list having pikachu, charmander, etc directly to it.

Hope this resolved your doubt.
Plz mark the doubt as resolved in my doubts section. :blush:

#3

I got this error ! my code ran , but it is giving just 49% accuracy for the test data. Is there anything wrong in my code, if not then what can i do to improve my code accuracy?
link for my updated code:-https://drive.google.com/file/d/1HBSn5K9BoIYgscNvMMaznPdhUnw2nY99/view?usp=sharing

#4

You loss curve contains many spikes, decrease learning rate and increase number of epochs. Do this until curve smoothens, and seems to reach its minima.

#5

I hope I’ve cleared your doubt. I ask you to please rate your experience here
Your feedback is very important. It helps us improve our platform and hence provide you
the learning experience you deserve.

On the off chance, you still have some questions or not find the answers satisfactory, you may reopen
the doubt.

#6

I got upto smoothening up my curves by decreasing learning rates ,but still i am able to achieve a maximum of 52% accuracy , is this ok ? or its still because of some mistake. I just want to ensure that 52% accuracy is genuine to be achieved using svm or it might be because of some other error?
Code link:-https://drive.google.com/file/d/1IbFlTgHtmqWr10dRXrkKb2HpaAuN05en/view?usp=sharing

#7

Hey @saksham_thukral, It will take some time. Wait for the reply, since i need to run the complete file.

#8

Hey @saksham_thukral, everything seems fine, now try to use sklearn’s svm classifier, with different kernels etc.

#9

I hope I’ve cleared your doubt. I ask you to please rate your experience here
Your feedback is very important. It helps us improve our platform and hence provide you
the learning experience you deserve.

On the off chance, you still have some questions or not find the answers satisfactory, you may reopen
the doubt.