Instructions to use sesame/csm-1b with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Transformers
How to use sesame/csm-1b with Transformers:
# Use a pipeline as a high-level helper from transformers import pipeline pipe = pipeline("text-to-speech", model="sesame/csm-1b")# Load model directly from transformers import AutoProcessor, AutoModelForTextToWaveform processor = AutoProcessor.from_pretrained("sesame/csm-1b") model = AutoModelForTextToWaveform.from_pretrained("sesame/csm-1b") - Notebooks
- Google Colab
- Kaggle
Prompting for pauses, hesitations etc.
#46
by matthen - opened
Is there any recommended way to write texts so that they are synthesized with pauses?
I'm finding that in general, randomly sampled speakers do not give significant pauses at sentence boundaries, at commas etc?
In my experience so far, using .. or ... will generate a pause. Of course that sometimes causes the opposite problem, and you end up with a pause that's too long.