ComfyUI Node
FL FishSpeech TTS
A ComfyUI node in 🐟FL FishSpeech with 10 inputs and 1 output.
FL FishSpeech TTS
- fs_model
- fs_reference
- audio
◄textHello, this is a test of the FishSpeech text to speech system.►
◄seed0►
◄temperature1.00►
◄top_p0.90►
◄top_k30►
◄repetition_penalty1.10►
◄chunk_length512►
◄max_new_tokens0►
Category🐟FL FishSpeech
Inputs (10)
| Name | Type | Default | Description |
|---|---|---|---|
| fs_model | FS_MODEL | — | |
| text | STRING | Hello, this is a test of the FishSpeech text to speech system. | Text to speak. Use <|speaker:X|> for multi-speaker, [laugh], [whispers] for emotion control. |
| fs_referenceopt | FS_REFERENCE | Optional reference audio encoding for voice cloning. | |
| seedopt | INT | 00–2147483647 | Random seed. 0 = random. |
| temperatureopt | FLOAT | 1.000.1–2 | Sampling temperature. Higher = more varied. v2.0 default is 1.0. |
| top_popt | FLOAT | 0.900.1–1 | Top-p nucleus sampling threshold. v2.0 default is 0.9. |
| top_kopt | INT | 301–100 | Top-k sampling. Limits to top K most likely tokens. |
| repetition_penaltyopt | FLOAT | 1.101–2 | Repetition penalty. Higher = less repetition. |
| chunk_lengthopt | INT | 51250–1000 | Max bytes per text chunk for iterative generation. |
| max_new_tokensopt | INT | 00–4096 | Maximum new tokens to generate per chunk. 0 = auto. |
Outputs (1)
| Name | Type | Description |
|---|---|---|
| audio | AUDIO | — |