TenserTensor
ee88f9be6c
Implement sampling nodes: - ksampler.py: standard sampler with basic parameters - ksampler_advanced.py: advanced sampler with step control and noise settings - ksampler_context.py: context-based sampler extracting parameters from workflow context - ksampler_two_stage.py: automated two-stage sampling (draft + refinement) Add markdown documentation for all sampler nodes
2.4 KiB
2.4 KiB
TT_KSamplerAdvanced
This description was generated with AI assistance. If you spot any errors, please create an issue on GitHub.
Advanced sampler with fine-grained control over denoising steps and noise addition.
Inputs
| Parameter | Type | Required | Description |
|---|---|---|---|
model |
MODEL | Yes | The diffusion model to use for sampling |
positive |
CONDITIONING | Yes | Positive conditioning (what to generate) |
negative |
CONDITIONING | Yes | Negative conditioning (what to avoid) |
latent |
LATENT | Yes | Latent image to denoise |
add_noise |
BOOLEAN | Yes | Add noise before sampling (default: True) |
full_denoise |
BOOLEAN | Yes | Complete full denoising (default: False) |
seed |
INT | Yes | Random seed for reproducibility (0 to 2^64-1, default: 0) |
steps |
INT | Yes | Total number of sampling steps (1-10000, default: 25) |
start_step |
INT | Yes | Step to start sampling from (0-10000, default: 0) |
last_step |
INT | Yes | Step to end sampling at (0-10000, default: 10000) |
cfg |
FLOAT | Yes | Classifier-free guidance scale (0.0-100.0, default: 1.5) |
sampler_name |
SAMPLER | Yes | Sampler algorithm to use |
scheduler |
SCHEDULER | Yes | Noise schedule to use |
denoise |
FLOAT | Yes | Denoising strength (0.0-1.0, default: 1.0) |
Outputs
| Parameter | Type | Description |
|---|---|---|
LATENT |
LATENT | Denoised latent image |
Usage
Advanced sampler for multi-stage workflows and fine control. Use start_step/last_step for partial denoising, add_noise to control noise injection, and denoise for strength control. Useful for img2img, refinement passes, and complex sampling workflows.