ComfyUI Node
Omnivoice TTS (Voice Design)
OmniVoice voice design via omnivoice-triton. Describe the target voice instead of providing reference audio.
Omnivoice TTS (Voice Design)
- audio
◄textHello, this is an OmniVoice designed voice test.►
◄instructfemale, young adult, high pitch►
◄runner_modehybrid►
◄num_step32►
◄guidance_scale2.0►
◄class_temperature0.00►
◄languageauto►
◄model_idk2-fsa/OmniVoice►
◄dtypefp16►
◄devicecuda►
CategoryOmnivoice-Triton
Inputs (10)
| Name | Type | Default | Description |
|---|---|---|---|
| text | STRING | Hello, this is an OmniVoice designed voice test. | — |
| instruct | STRING | female, young adult, high pitch | — |
| runner_mode | COMBO | hybrid | 4 options: base, triton, faster, hybrid |
| num_stepopt | INT | 324–64 | — |
| guidance_scaleopt | FLOAT | 2.00–10 | — |
| class_temperatureopt | FLOAT | 0.000–2 | — |
| languageopt | COMBO | auto | 4 options: auto, ko, en, zh |
| model_idopt | STRING | k2-fsa/OmniVoice | — |
| dtypeopt | COMBO | fp16 | 3 options: fp16, bf16, fp32 |
| deviceopt | COMBO | cuda | 2 options: cuda, cpu |
Outputs (1)
| Name | Type | Description |
|---|---|---|
| audio | AUDIO | — |