An Autoencoder for WAV files
- Create a folder named 'audio_wav' and a folder named 'output'
- Put all WAV files for training in the audio_wav folder
- Open up your terminal within the folder and run
python3 encoder.py
-
If you're only looking at a single batch of songs, it'd be wise to move the
ch1_song, ch2_song, sample_rate = next_batch(i, batch_size, sess)call outside of the training loops. This way it's only called once. -
If you find yourself running out of memory too quickly, reduce the songs_per_batch, and the node sizes for the input layers.