I use the following repo. to train a TTS on my own dataset : Tacotron 2
All stuffs are going well, but when I put
distributed_run = True in
hparams.py the process has been stucked at
init_process_group function in
train.py at line 35 (I put my
world_size = 3). My GPU driver and cuda are okey! Is this a bug? If yes how can I fix it?