Varun K. Srikanth
Varun K. Srikanth
Hey, has anyone found a solution for this?
The trouble seems to be from autoregressive_batch_size. Set a value and play around with it - it should then run! (autoregressive_batch_size=15 works well on my setup) tortoise-tts/tortoise/api.py `class TextToSpeech:` def...
was facing the same problem. After execution the utilized memory still persists (~11GB). Deleting the generated sample or tts or any other variable doesn't clear it. Workaround was using multiprocessing...
Hey @theveloped, Did you manage to find a solution for this?