i have my own training data which is excatly in the format you are using.16 mono 16khz pcm
and the csv file is in shape wav filename, size,transcript. i changed the shell script accordingly
and the training is happening. I have a doubt here is it using pre trained model or creating a new model with my data? If it is creating a new model then how can i use pretrained model with my data.
Thanks in advance.!