| { |
| "1": { |
| "class_type": "UNETLoader", |
| "inputs": { |
| "unet_name": "auk_flash.safetensors", |
| "weight_dtype": "default" |
| }, |
| "_meta": { |
| "title": "AuK-Flash transformer" |
| } |
| }, |
| "2": { |
| "class_type": "CLIPLoader", |
| "inputs": { |
| "clip_name": "auk_qwen25_omni_3b.safetensors", |
| "type": "auk", |
| "device": "default" |
| }, |
| "_meta": { |
| "title": "AuK Qwen conditioning" |
| } |
| }, |
| "3": { |
| "class_type": "VAELoader", |
| "inputs": { |
| "vae_name": "auk_audio_vae.safetensors" |
| }, |
| "_meta": { |
| "title": "AuK audio VAE" |
| } |
| }, |
| "4": { |
| "class_type": "PrimitiveStringMultiline", |
| "inputs": { |
| "value": "Say the following with the same voice: \"Hello there. This is a short audio test.\"" |
| }, |
| "_meta": { |
| "title": "Instruction" |
| } |
| }, |
| "15": { |
| "class_type": "LoadAudio", |
| "inputs": { |
| "audio": "reference.wav" |
| }, |
| "_meta": { |
| "title": "Choose reference audio" |
| } |
| }, |
| "5": { |
| "class_type": "PrimitiveFloat", |
| "inputs": { |
| "value": 5.0 |
| }, |
| "_meta": { |
| "title": "Target duration (seconds)" |
| } |
| }, |
| "9": { |
| "class_type": "RandomNoise", |
| "inputs": { |
| "noise_seed": 42 |
| }, |
| "_meta": { |
| "title": "Initial noise seed" |
| } |
| }, |
| "10": { |
| "class_type": "KSamplerSelect", |
| "inputs": { |
| "sampler_name": "euler" |
| }, |
| "_meta": { |
| "title": "Four-step Euler" |
| } |
| }, |
| "11": { |
| "class_type": "ManualSigmas", |
| "inputs": { |
| "sigmas": "1.0, 0.9238795042037964, 0.7071067690849304, 0.3826833963394165, 0.0" |
| }, |
| "_meta": { |
| "title": "Exact AuK-Flash schedule" |
| } |
| }, |
| "6": { |
| "class_type": "AuKConditioning", |
| "inputs": { |
| "reference_seed": 0, |
| "clip": [ |
| "2", |
| 0 |
| ], |
| "reference_audio": [ |
| "15", |
| 0 |
| ], |
| "vae": [ |
| "3", |
| 0 |
| ], |
| "instruction": [ |
| "4", |
| 0 |
| ] |
| }, |
| "_meta": { |
| "title": "Encode instruction and reference" |
| } |
| }, |
| "7": { |
| "class_type": "EmptyAuKLatentAudio", |
| "inputs": { |
| "batch_size": 1, |
| "seconds": [ |
| "5", |
| 0 |
| ] |
| }, |
| "_meta": { |
| "title": "24 kHz \u00b7 50 latent frames/sec" |
| } |
| }, |
| "8": { |
| "class_type": "BasicGuider", |
| "inputs": { |
| "model": [ |
| "1", |
| 0 |
| ], |
| "conditioning": [ |
| "6", |
| 0 |
| ] |
| }, |
| "_meta": { |
| "title": "Conditional only \u00b7 no extra CFG" |
| } |
| }, |
| "12": { |
| "class_type": "SamplerCustomAdvanced", |
| "inputs": { |
| "noise": [ |
| "9", |
| 0 |
| ], |
| "guider": [ |
| "8", |
| 0 |
| ], |
| "sampler": [ |
| "10", |
| 0 |
| ], |
| "sigmas": [ |
| "11", |
| 0 |
| ], |
| "latent_image": [ |
| "7", |
| 0 |
| ] |
| }, |
| "_meta": { |
| "title": "Generate audio latents" |
| } |
| }, |
| "13": { |
| "class_type": "VAEDecodeAudio", |
| "inputs": { |
| "normalize": false, |
| "samples": [ |
| "12", |
| 0 |
| ], |
| "vae": [ |
| "3", |
| 0 |
| ] |
| }, |
| "_meta": { |
| "title": "Decode \u00b7 preserve waveform amplitude" |
| } |
| }, |
| "14": { |
| "class_type": "SaveAudioAdvanced", |
| "inputs": { |
| "filename_prefix": "audio/auk_reference", |
| "format": "flac", |
| "audio": [ |
| "13", |
| 0 |
| ] |
| }, |
| "_meta": { |
| "title": "Save AuK audio" |
| } |
| } |
| } |
|
|
