ComfyUI Node
OmniVoice Load Model
A ComfyUI node in OmniVoice/Load with 6 inputs and 1 output.
OmniVoice Load Model
- reference audio (optional)
- pipe
◄deviceauto►
◄dtypeauto►
◄offload_after_generatefalse►
◄asr_model_nameopenai/whisper-large-v3-turbo►
◄transcription override—►
CategoryOmniVoice/Load
Inputs (6)
| Name | Type | Default | Description |
|---|---|---|---|
| device | COMBO | auto | Model execution device |
| dtype | COMBO | auto | Model tensor dtype |
| offload_after_generate | BOOLEAN | false | Unload OmniVoice from memory after generation |
| asr_model_name | COMBO | openai/whisper-large-v3-turbo | Whisper model used when transcription override is empty |
| reference audio (optional)opt | AUDIO | Reference speaker audio for voice clone | |
| transcription overrideopt | STRING | Optional transcript for reference audio; ignored if no reference audio |
Outputs (1)
| Name | Type | Description |
|---|---|---|
| pipe | OMNIVOICE_PIPE | — |