pip install for only converting text to speech? #4032
-
I'm on Debian and tried pip install TTS. About 80 packages installed fine. After that I see that it wants to install a lot of large nvidia packages. Do I need those? I'm trying to only convert text to speech, no creating or training voices. Is there a way to install only what I need for my purpose? |
Beta Was this translation helpful? Give feedback.
Replies: 1 comment 2 replies
-
The Nvidia/CUDA packages are also needed for inference on a GPU, which is much faster than on CPU, so you might still want to install them. Otherwise you could also install just the CPU-version of Pytorch with their instructions: https://pytorch.org/get-started/locally/ I'd also suggest to install our fork instead (with |
Beta Was this translation helpful? Give feedback.
They are all transitive dependencies that come through Pytorch, so there is no list in Coqui, but Pytorch 2.4.1 has the following dependencies for me: