Closed inchpunch closed 5 years ago
What data set have you used for evaluation?
I used librivox-dev-clean, the default dataset in the example_config for small model.
I got 32% after training. Maybe it does not have any problem due to the small network and dataset compared with the original DS2 and Librispeech.
It looks so. BTW that was from default "num_epochs"=12. I ran 28 epochs and then got 25.2%.
I ran DS2 small model in evaluation mode without using language model. WER is around 29%. I saw in the default base_params num_epochs is 12. Is this performance expected? Thanks!