How much data for deepspeech for japanese english accent

[Vinayak_Sable]

I need to train the model on the japanese language with an english accent
some question about that _:

1. anybody already worked on it?
2. how much data will be required?
3. training time
4. comparison accuracy(if first yes)