tinyKSampler¶
Documentation¶
- Class name:
ttN KSampler_v2 - Category:
🌏 tinyterra/base - Output node:
True
The ttN KSampler_v2 node is designed for advanced sampling in generative models, offering a flexible interface to control the sampling process. It integrates various sampling and scheduling strategies, noise control mechanisms, and optional conditioning inputs to generate or refine images based on latent representations. This node is tailored for customization and experimentation in image generation workflows, enabling users to fine-tune the balance between fidelity and diversity in the generated outputs.
Input types¶
Required¶
model- Specifies the generative model used for sampling, central to the generation process.
- Comfy dtype:
MODEL - Python dtype:
object
positive- Optional positive conditioning to guide the generation towards desired attributes or content.
- Comfy dtype:
CONDITIONING - Python dtype:
object
negative- Optional negative conditioning to steer the generation away from undesired attributes or content.
- Comfy dtype:
CONDITIONING - Python dtype:
object
latent- unknown
- Comfy dtype:
LATENT - Python dtype:
unknown
vae- unknown
- Comfy dtype:
VAE - Python dtype:
unknown
lora_name- unknown
- Comfy dtype:
COMBO[STRING] - Python dtype:
unknown
lora_strength- unknown
- Comfy dtype:
FLOAT - Python dtype:
unknown
upscale_method- unknown
- Comfy dtype:
COMBO[STRING] - Python dtype:
unknown
upscale_model_name- unknown
- Comfy dtype:
COMBO[STRING] - Python dtype:
unknown
factor- unknown
- Comfy dtype:
FLOAT - Python dtype:
unknown
rescale- unknown
- Comfy dtype:
COMBO[STRING] - Python dtype:
unknown
percent- unknown
- Comfy dtype:
INT - Python dtype:
unknown
width- unknown
- Comfy dtype:
INT - Python dtype:
unknown
height- unknown
- Comfy dtype:
INT - Python dtype:
unknown
longer_side- unknown
- Comfy dtype:
INT - Python dtype:
unknown
crop- unknown
- Comfy dtype:
COMBO[STRING] - Python dtype:
unknown
steps- unknown
- Comfy dtype:
INT - Python dtype:
unknown
cfg- Sets the conditioning-free guidance (CFG) scale, adjusting the influence of conditional inputs on the generation.
- Comfy dtype:
FLOAT - Python dtype:
float
sampler_name- Selects the specific sampling algorithm to be used, allowing for customization of the sampling behavior.
- Comfy dtype:
COMBO[STRING] - Python dtype:
str
scheduler- Chooses the scheduling algorithm for controlling the sampling process, impacting the progression of noise reduction.
- Comfy dtype:
COMBO[STRING] - Python dtype:
str
denoise- Determines the level of denoising applied to the final output, affecting the clarity and quality of the generated image.
- Comfy dtype:
FLOAT - Python dtype:
float
image_output- unknown
- Comfy dtype:
COMBO[STRING] - Python dtype:
unknown
save_prefix- unknown
- Comfy dtype:
STRING - Python dtype:
unknown
file_type- unknown
- Comfy dtype:
COMBO[STRING] - Python dtype:
unknown
embed_workflow- unknown
- Comfy dtype:
BOOLEAN - Python dtype:
unknown
Optional¶
clip- unknown
- Comfy dtype:
CLIP - Python dtype:
unknown
seed- unknown
- Comfy dtype:
INT - Python dtype:
unknown
input_image_override- unknown
- Comfy dtype:
IMAGE - Python dtype:
unknown
adv_xyPlot- unknown
- Comfy dtype:
ADV_XYPLOT - Python dtype:
unknown
Output types¶
model- Comfy dtype:
MODEL - The generative model used for sampling, central to the generation process.
- Python dtype:
object
- Comfy dtype:
positive- Comfy dtype:
CONDITIONING - The positive conditioning applied to guide the generation towards desired attributes or content.
- Python dtype:
object
- Comfy dtype:
negative- Comfy dtype:
CONDITIONING - The negative conditioning applied to steer the generation away from undesired attributes or content.
- Python dtype:
object
- Comfy dtype:
latent- Comfy dtype:
LATENT - The latent representation used or generated during the sampling process.
- Python dtype:
object
- Comfy dtype:
vae- Comfy dtype:
VAE - The optional variational autoencoder (VAE) model used in the generation process, enhancing the diversity of outputs.
- Python dtype:
object
- Comfy dtype:
clip- Comfy dtype:
CLIP - The optional CLIP model included to guide the generation process based on textual descriptions.
- Python dtype:
object
- Comfy dtype:
images- Comfy dtype:
IMAGE - The generated or refined images based on the provided inputs and settings, showcasing the capabilities of the sampling process.
- Python dtype:
object
- Comfy dtype:
seed- Comfy dtype:
INT - The seed used for noise generation, ensuring reproducibility of the randomness.
- Python dtype:
int
- Comfy dtype:
plot_image- Comfy dtype:
IMAGE - Visual representation of the generation process or outcomes, if applicable.
- Python dtype:
object
- Comfy dtype:
Usage tips¶
- Infra type:
GPU - Common nodes: unknown
Source code¶
class ttN_KSampler_v2:
version = '2.3.1'
def __init__(self):
pass
@classmethod
def INPUT_TYPES(cls):
return {"required": {
"model": ("MODEL",),
"positive": ("CONDITIONING",),
"negative": ("CONDITIONING",),
"latent": ("LATENT",),
"vae": ("VAE",),
"lora_name": (["None"] + folder_paths.get_filename_list("loras"),),
"lora_strength": ("FLOAT", {"default": 1.0, "min": -10.0, "max": 10.0, "step": 0.01}),
"upscale_method": (UPSCALE_METHODS, {"default": "None"}),
"upscale_model_name": (UPSCALE_MODELS,),
"factor": ("FLOAT", {"default": 2, "min": 0.0, "max": 10.0, "step": 0.25}),
"rescale": (["by percentage", "to Width/Height", 'to longer side - maintain aspect', 'None'],),
"percent": ("INT", {"default": 50, "min": 0, "max": 1000, "step": 1}),
"width": ("INT", {"default": 1024, "min": 64, "max": MAX_RESOLUTION, "step": 8}),
"height": ("INT", {"default": 1024, "min": 64, "max": MAX_RESOLUTION, "step": 8}),
"longer_side": ("INT", {"default": 1024, "min": 64, "max": MAX_RESOLUTION, "step": 8}),
"crop": (CROP_METHODS,),
"steps": ("INT", {"default": 20, "min": 1, "max": 10000}),
"cfg": ("FLOAT", {"default": 8.0, "min": 0.0, "max": 100.0}),
"sampler_name": (comfy.samplers.KSampler.SAMPLERS,),
"scheduler": (comfy.samplers.KSampler.SCHEDULERS + CUSTOM_SCHEDULERS,),
"denoise": ("FLOAT", {"default": 1.0, "min": 0.0, "max": 1.0, "step": 0.01}),
"image_output": (["Hide", "Preview", "Save", "Hide/Save", "Disabled"],),
"save_prefix": ("STRING", {"default": "ComfyUI"}),
"file_type": (OUTPUT_FILETYPES,{"default": "png"}),
"embed_workflow": ("BOOLEAN", {"default": True}),
},
"optional": {
"clip": ("CLIP",),
"seed": ("INT", {"default": 0, "min": 0, "max": 0xffffffffffffffff}),
"input_image_override": ("IMAGE",),
"adv_xyPlot": ("ADV_XYPLOT",),
},
"hidden": {
"prompt": "PROMPT", "extra_pnginfo": "EXTRA_PNGINFO", "my_unique_id": "UNIQUE_ID",
"ttNnodeVersion": ttN_KSampler_v2.version
},
}
RETURN_TYPES = ("MODEL", "CONDITIONING", "CONDITIONING", "LATENT", "VAE", "CLIP", "IMAGE", "INT", "IMAGE")
RETURN_NAMES = ("model", "positive", "negative", "latent","vae", "clip", "images", "seed", "plot_image")
OUTPUT_NODE = True
FUNCTION = "sample"
CATEGORY = "🌏 tinyterra/base"
def sample(self, model, positive, negative, latent, vae,
lora_name, lora_strength,
steps, cfg, sampler_name, scheduler, image_output, save_prefix, file_type, embed_workflow, denoise=1.0,
input_image_override=None,
clip=None, seed=None, adv_xyPlot=None, upscale_model_name=None, upscale_method=None, factor=None, rescale=None, percent=None, width=None, height=None, longer_side=None, crop=None,
prompt=None, extra_pnginfo=None, my_unique_id=None, start_step=None, last_step=None, force_full_denoise=False, disable_noise=False):
my_unique_id = int(my_unique_id)
ttN_save = ttNsave(my_unique_id, prompt, extra_pnginfo)
def process_sample_state(model, images, clip, samples, vae, seed, positive, negative, lora_name, lora_model_strength, lora_clip_strength,
upscale_model_name, upscale_method, factor, rescale, percent, width, height, longer_side, crop,
steps, cfg, sampler_name, scheduler, denoise,
image_output, save_prefix, file_type, embed_workflow, prompt, extra_pnginfo, my_unique_id, preview_latent, start_step=start_step, last_step=last_step, force_full_denoise=force_full_denoise, disable_noise=disable_noise):
# Load Lora
if lora_name not in (None, "None"):
if clip == None:
raise ValueError(f"tinyKSampler [{my_unique_id}] - Lora requires CLIP model")
model, clip = loader.load_lora(lora_name, model, clip, lora_model_strength, lora_clip_strength)
# Upscale samples if enabled
if upscale_method != "None":
samples = sampler.handle_upscale(samples, upscale_method, factor, crop, upscale_model_name, vae, images, rescale, percent, width, height, longer_side)
samples = sampler.common_ksampler(model, seed, steps, cfg, sampler_name, scheduler, positive, negative, samples, denoise=denoise, preview_latent=preview_latent, start_step=start_step, last_step=last_step, force_full_denoise=force_full_denoise, disable_noise=disable_noise)
results = list()
if (image_output != "Disabled"):
# Save images
latent = samples["samples"]
images = vae.decode(latent)
results = ttN_save.images(images, save_prefix, image_output, embed_workflow, file_type)
if image_output in ("Hide", "Hide/Save", "Disabled"):
return (model, positive, negative, samples, vae, clip, images, seed, None)
return {"ui": {"images": results},
"result": (model, positive, negative, samples, vae, clip, images, seed, None)}
def process_xyPlot(model, clip, samp_samples, vae, seed, positive, negative, lora_name, lora_model_strength, lora_clip_strength,
steps, cfg, sampler_name, scheduler, denoise,
image_output, save_prefix, file_type, embed_workflow, prompt, extra_pnginfo, my_unique_id, preview_latent, adv_xyPlot):
random.seed(seed)
executor = xyExecutor()
plotter = ttNadv_xyPlot(adv_xyPlot, my_unique_id, prompt, extra_pnginfo, save_prefix, image_output, executor)
plot_image, images, samples = plotter.xy_plot_process()
plotter.reset()
del executor, plotter
if samples is None and images is None:
return process_sample_state(model, images, clip, samp_samples, vae, seed, positive, negative, lora_name, lora_model_strength, lora_clip_strength,
upscale_model_name, upscale_method, factor, rescale, percent, width, height, longer_side, crop,
steps, cfg, sampler_name, scheduler, denoise,
image_output, save_prefix, file_type, embed_workflow, prompt, extra_pnginfo, my_unique_id, preview_latent, start_step=start_step, last_step=last_step, force_full_denoise=force_full_denoise, disable_noise=disable_noise)
plot_result = ttN_save.images(plot_image, save_prefix, image_output, embed_workflow, file_type)
#plot_result.extend(ui_results)
if image_output in ("Hide", "Hide/Save"):
return (model, positive, negative, samples, vae, clip, images, seed, plot_image)
return {"ui": {"images": plot_result}, "result": (model, positive, negative, samples, vae, clip, images, seed, plot_image)}
preview_latent = True
if image_output in ("Hide", "Hide/Save", "Disabled"):
preview_latent = False
if adv_xyPlot is None:
return process_sample_state(model, input_image_override, clip, latent, vae, seed, positive, negative, lora_name, lora_strength, lora_strength,
upscale_model_name, upscale_method, factor, rescale, percent, width, height, longer_side, crop,
steps, cfg, sampler_name, scheduler, denoise, image_output, save_prefix, file_type, embed_workflow, prompt, extra_pnginfo, my_unique_id, preview_latent)
else:
return process_xyPlot(model, clip, latent, vae, seed, positive, negative, lora_name, lora_strength, lora_strength, steps, cfg, sampler_name,
scheduler, denoise, image_output, save_prefix, file_type, embed_workflow, prompt, extra_pnginfo, my_unique_id, preview_latent, adv_xyPlot)