Closed DevZiegler closed 4 years ago
Hey! Thank you. Unfortunately, the experiments still are in progress. I hope it will change soon. Rafal
Hello, Thanks a lot for a great piece of code. I guess that pretrained model still doesn't exist so do I have to train it by myself?
hey ! Thank you, unfortunately still no... we work on
@rolczynski Do you recommend any Polish (or even English) free dataset to train your model? (which doesn't require a lot of modification e.g. wave files up to 5 sec)
@AleksandraPestka I highly recommend https://github.com/danijel3/ClarinStudioKaldi, and you can contact with @danijel3. We will publish here the initial version of the Polish model at the beginning of September
@AleksandraPestka, we've just hosted an initial pre-trained model, see instructions here: https://github.com/rolczynski/DeepSpeech-Keras/tree/master/models
Could you guide me how to test this model with pre-trained weights and .wav files provided in test/data/audio, please? I've tried to follow the code in readme, but I've got stuck with the folders hierarchy. To be precise - which script should I run to test this model and from which folder? ( in order to avoid errors)
@AleksandraPestka, we can arrange the 1hour session and together we will explore how it works, but... one thing! Then you write README, right?
Are you asking me about updating README? As a nonauthor of this project I think I'm not supposed to do that... but if I can still help with anything, I will
Is it still possible to get help to test pre-trained model?
Hi, first of all, thanks for the work and the project. Are there any pretrained models for your implementation? And if so, where can I find them? thx