how do i use this repo?

[akshat9425]

hi,

i want to create voice authentication and for the same i need to recognize invalid speakers please tell me how could i use this repo

waiting for your valuable reply thanks in advance

[jan25]

This repo was one of our projects back in college. It uses simple supervised learning and was trained on only handful of speakers. I don't believe this repo as is can scale for your requirements where i assume you will have many more users to authenticate. However, if you job is just to differentiate valid and invalid users, you could change the output classes to 0/1, instead of one class for each user as in this repo, and retrain+fine-tune the model according to your needs.

[akshat9425]

Thanks for reply for your repo you used this line : from features import mfcc i installed features library but there is no file of mfcc present inside it please tell me what to do

[akshat9425]

I replace features from python_speech_features and issue is resolved to extract mfcc but will you please please send me a sample command to run python main.py

i am using this command python main.py -train /home/user/invalid/Speaker-Recognition/train_data/ -test /home/user/invalid/Speaker-Recognition/test_data/ -recog /home/user/invalid/Speaker-Recognition/recog_speaker/

and getting this issue ValueError: invalid literal for int() with base 10: '-test'

[gitrohini]

I also getting same problem but then I try this python main.py -train /home/user/18_10/spkerRecog3/Speaker-Recognition/train_data/ it works