ComfyUI Node
WavTTS Load Model
Load WavTTS with ComfyUI memory tracking. Official weights are stored under ComfyUI/models/wavtts/.
WavTTS Load Model
- wavtts_model
◄checkpointWavTTS 16k FP32 - drbaph/WavTTS (auto-download)►
◄dtypeauto►
◄deviceauto►
◄attentionauto►
◄download_if_missingtrue►
CategoryWavTTS
Inputs (5)
| Name | Type | Default | Description |
|---|---|---|---|
| checkpoint | COMBO | WavTTS 16k FP32 - drbaph/WavTTS (auto-download) | Cataloged model weights from drbaph/WavTTS. FP32 is stable; mixed BF16 is the low-VRAM auto-download path. |
| dtype | COMBO | auto | Weight dtype. auto keeps the selected checkpoint dtype. FP32 is the stable WavTTS inference path. |
| device | COMBO | auto | Device for WavTTS. auto uses ComfyUI's current torch device. |
| attention | COMBO | auto | auto resolves to sdpa. sdpa uses upstream WavTTS torch SDPA. flash_attention uses upstream flash_attn. sageattention applies a WavTTS SDPA patch during generation. |
| download_if_missing | BOOLEAN | true | Download the selected preset into ComfyUI/models/wavtts/ when it is missing. |
Outputs (1)
| Name | Type | Description |
|---|---|---|
| wavtts_model | WAVTTS_MODEL | — |