search

HPI-DeepLearning / crnn-lid

Code for the paper Language Identification Using Deep Convolutional Recurrent Neural Networks
GNU General Public License v3.0
105 stars 48 forks source link

Training steps. #21

Closed Varuzhan97 closed 4 years ago

Varuzhan97 commented 4 years ago

Can anyone explaint me training process step by step.I have tried every method and way but no results only errors:).

Bartzi commented 4 years ago

Well, you first download the data with all tools in the data directory, then you create a config.yaml based on our template, and then you start the training. It is as easy as that.

Varuzhan97 commented 4 years ago

Yes I have done. Program shows that download is in process but there in no any downloaded data in folder.

Bartzi commented 4 years ago

This is odd :shrug: are you using YouTube download? In the end the question is: what are you exactly trying?

Varuzhan97 commented 4 years ago

I am trying to download data from youtube using download_youtube.py.So what must create download_youtube after downloading data?

Bartzi commented 4 years ago

your output directory should contain some subdirectories, such as raw and in there you should find .wav files. and you should also have a .csv file

Varuzhan97 commented 4 years ago

After downloading data it creates train test validation csv files but this files are empty.

Bartzi commented 4 years ago

hmm, did the script provide some output which videos it tried to download? Seems that the videos are not found.

Varuzhan97 commented 4 years ago

Yes.Look my subdirectory contains raw folder with .wav file . after downloading data and finishing download .csv files are still empty

Varuzhan97 commented 4 years ago

must i create spectograms after downloading data?