network_bending

AudioFeatureExtractor

Extract audio features for guided manipulation. - Features: Spectral centroid, rolloff, zero-crossing rate, MFCC, bandwidth, RMS energy, tempo, onset strength - Use: Guide latent manipulations based on audio characteristics

Audio/Analysis · 1 input · 3 parameters · 1 output

AudioLatentBlend

Blend multiple audio latents with various blend modes. - Modes: Add, multiply, screen, overlay, soft light, difference, exclusion - Supports: Up to 4 audio inputs with individual weights

Audio/Latent · 4 inputs · 7 parameters · 1 output

AudioLatentGuidance

Guide the diffusion process with reference latents. - Modes: Additive, multiplicative, attention, gradient - Control: Start/end steps, guidance strength

Audio/Conditioning · 3 inputs · 4 parameters · 1 output

AudioLatentInterpolate

Smoothly interpolate between two audio latents. - Modes: Linear, spherical (SLERP), cubic, sine - Use cases: Morphing between sounds, creating transitions

Audio/Latent · 2 inputs · 3 parameters · 1 output

AudioLatentManipulator

Apply various transformations to audio latents. - Operations: - Frequency shift: Shift frequency content - Temporal stretch: Time-stretch without pitch change - Harmonic emphasis: Enhance harmonic content - Noise injection: Add controlled noise - Dynamic range: Compress/expand dynamics - Spatial transform: Rotate in latent space - Resonance filter: Apply resonant filtering

Audio/Latent · 2 inputs · 4 parameters · 1 output

AudioReferenceEncoder

Encode multiple audio references into a combined representation. - Combination modes: Average, weighted, PCA, attention - Features: Segment extraction, weighted blending

Audio/Reference · 5 inputs · 7 parameters · 2 outputs

AudioStyleTransfer

Transfer style characteristics between audio sources. - Modes: - Global: Statistics matching - Frequency bands: Band-wise style transfer - Temporal segments: Time-based style transfer - Adaptive: Correlation-based transfer - Neural style: Deep feature matching

Audio/Style · 2 inputs · 5 parameters · 1 output

AudioVAEDecode

Decodes audio latents back to waveforms. - Inputs: Audio latent, VAE model - Parameters: Denormalize (bool) - Outputs: Audio waveform

Audio/Latent · 2 inputs · 1 parameter · 1 output

AudioVAEEncode

Encodes audio waveforms into Stable Audio's latent space. - Inputs: Audio waveform, VAE model - Parameters: Normalize (bool), target sample rate - Outputs: Audio latent, latent info

Audio/Latent · 2 inputs · 2 parameters · 2 outputs

InvertedPruning

Inverted Model Pruning - Selectively removes critical weights for artistic degradation Instead of preserving important weights (standard pruning), this node removes them, creating unique artistic effects. Based on the inverted Lottery Ticket Hypothesis.

Network Bending · 1 input · 8 parameters · 1 output

LatentFormatConverter

Convert between audio and image latent formats This node handles conversion between different latent tensor shapes: - Audio latents: [batch, channels, length] (3D) - Image latents: [batch, channels, height, width] (4D)

Network Bending · 1 input · 6 parameters · 1 output

max

Compact description unavailable for this node.

ModelMixer

Mix two models together with various blending modes

Network Bending · 2 inputs · 2 parameters · 1 output

multiline

Compact description unavailable for this node.

NetworkBending

Network Bending Node - Performs various modifications on neural network models This node allows for creative manipulation of loaded model checkpoints including: - Weight noise injection - Layer freezing/unfreezing - Weight scaling - Activation function replacement - Layer pruning - Weight mixing between models

Network Bending · 2 inputs · 5 parameters · 1 output

NetworkBendingAdvanced

Advanced Network Bending Node - Extended operations for model manipulation

Network Bending · 1 input · 3 parameters · 1 output

VAEChannelManipulator

Advanced channel manipulation for VAE latents

Network Bending/Vae · 2 inputs · 3 parameters · 1 output

VAELatentBending

Perform network bending operations directly on VAE latent streams

Network Bending/Vae · 1 input · 4 parameters · 1 output

VAEMixer

Mix two VAE models together with various blending strategies

Network Bending/Vae · 2 inputs · 2 parameters · 1 output

VAENetworkBending

VAE Network Bending - Performs creative modifications on VAE models This node allows manipulation of VAE encoder/decoder networks including: - Weight noise injection in encoder/decoder - Asymmetric modifications (encoder vs decoder) - Latent space dimension manipulation - Channel-specific modifications

Network Bending/Vae · 1 input · 5 parameters · 1 output