Ok, nevermind; yep, it was easy 
After loading TTS model I received such an error:
I am using this released model, however, I wonder, where I should do such an adjustment in the shape of current model?
I am using this released model, however, I wonder, where I should do such an adjustment in the shape of current model?
You probably have used another config with tacotron as model maybe? Check and let me know.
I checked both before running the notebook and I fixed it. Maybe I should remove gradual_training field from the original config.json, but it shouldn’t affect the work, since it works only to Tacotron, not Tacotron2.
I’ll send you the config file I trained it with tomorrow. Try it with that. I am not sure if the model has been changed since the branch I’ve used to train this. I can’t seem to remember any other changes.
cool 
so, waiting for config then