search

AliaksandrSiarohin / pose-gan

381 stars 78 forks source link

Parameters to enhance the training speed #45

Closed gyawali9 closed 4 years ago

gyawali9 commented 4 years ago

what are the required parameters to be changed to train our self made data in an efficient way to enhance the accuracy as well?

AliaksandrSiarohin commented 4 years ago

Not clear, what you mean.

gyawali9 commented 4 years ago

@AliaksandrSiarohin I tried to train 500 images with this code. Even for a single epoch it takes more training time. So,how to minimize the training time? What should be the value of parameters so that it will take less time to train my data?

AliaksandrSiarohin commented 4 years ago

This depends on your data. What is the resolution of your data? How many pairs you have? Single epoch here is simply 1000 iterations, you have the same? Slowe compared to what dataset market or fashion?

gyawali9 commented 4 years ago

thank you sir

gyawali9 commented 4 years ago

thank u sir

On Wed, 13 Nov 2019 at 19:41, AliaksandrSiarohin notifications@github.com wrote:

This depends on your data. What is the resolution of your data? How many pairs you have? Single epoch here is simply 1000 iterations, you have the same? Slowe compared to what dataset market or fashion?

— You are receiving this because you modified the open/close state. Reply to this email directly, view it on GitHub https://github.com/AliaksandrSiarohin/pose-gan/issues/45?email_source=notifications&email_token=AJTUSTM26L262SA2VBFZFVDQTQBRVA5CNFSM4JMSQMUKYY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGOED6HAYY#issuecomment-553414755, or unsubscribe https://github.com/notifications/unsubscribe-auth/AJTUSTLRYPLBEY5TXXLDBIDQTQBRVANCNFSM4JMSQMUA .