Fantastic implementation, thank you for sharing it with the community.
I have a problem with the node cache. If I connect a reference_audio input to the 🎤 ChatterBox Voice TTS node, it correctly one-shot clones the voice and generates speech.
However, if I later disconnect the reference_audio input and re-run the workflow, the node doesn't revert back to using the default voice (which I would prefer), but keeps using the last used reference audio.
The only way for me to revert to the default voice is by either pressing the Free model and node cache button in the ComfyUI top menu bar, or restart the system.
Any chance that cache can be ignored?
Thank you for considering this request.
Fantastic implementation, thank you for sharing it with the community.
I have a problem with the node cache. If I connect a
reference_audioinput to the🎤 ChatterBox Voice TTSnode, it correctly one-shot clones the voice and generates speech.However, if I later disconnect the
reference_audioinput and re-run the workflow, the node doesn't revert back to using the default voice (which I would prefer), but keeps using the last used reference audio.The only way for me to revert to the default voice is by either pressing the
Free model and node cachebutton in the ComfyUI top menu bar, or restart the system.Any chance that cache can be ignored?
Thank you for considering this request.