I am running DeepSpeech (0.3) on google colab. My whole dataset consists of 30000 files
and training set contains ~22000 files
. The preprocessing step is taking a lot time. Preprocessing the training set took 2+ hours
and still did not complete.
The training file specifications are:
Sample Rate: 16000
Channel: 1
Encoding: 16 bit signed Integer PCM
Dataset folder structure is as follows:
dataset
wav
audio_file1.wav
audio_file2.wav
...............
Previously, I trained with ~1500 audio
files and Preprocessing completed in seconds. So, I can’t figure out the issue here. What might be the issue other than colab?