21 KiB
21 KiB
Documentation
- Class name: ttN pipeKSamplerSDXL_v2
- Category: ttN/pipe
- Output node: True
ttN_pipeKSamplerSDXL_v2节点旨在促进高级图像采样过程，融合了各种技术和参数以增强和定制图像生成。它利用了一种复杂的采样算法，允许对图像生成过程进行详细控制，包括噪声管理、去噪以及应用LoRA调整来微调模型。
Input types
Required
- sdxl_pipe
- 未知
- Comfy dtype: PIPE_LINE_SDXL
- Python dtype: unknown
- lora_name
- 指定要调整的LoRA参数的名称，在微调模型行为和输出质量方面起着关键作用。
- Comfy dtype: COMBO[STRING]
- Python dtype: str
- lora_strength
- 决定LoRA调整的强度，通过改变模型的内部表示直接影响模型的输出。
- Comfy dtype: FLOAT
- Python dtype: float
- upscale_method
- 指定用于放大生成图像的方法，影响图像的分辨率和清晰度。
- Comfy dtype: COMBO[STRING]
- Python dtype: str
- upscale_model_name
- 未知
- Comfy dtype: COMBO[STRING]
- Python dtype: unknown
- factor
- 决定图像放大的倍数，直接影响最终图像的大小和细节。
- Comfy dtype: FLOAT
- Python dtype: float
- rescale
- 未知
- Comfy dtype: COMBO[STRING]
- Python dtype: unknown
- percent
- 未知
- Comfy dtype: INT
- Python dtype: unknown
- width
- 未知
- Comfy dtype: INT
- Python dtype: unknown
- height
- 未知
- Comfy dtype: INT
- Python dtype: unknown
- longer_side
- 未知
- Comfy dtype: INT
- Python dtype: unknown
- crop
- 未知
- Comfy dtype: COMBO[STRING]
- Python dtype: unknown
- sampler_state
- 控制采样器的状态，影响采样过程和最终的图像生成。
- Comfy dtype: COMBO[STRING]
- Python dtype: str
- base_steps
- 未知
- Comfy dtype: INT
- Python dtype: unknown
- cfg
- 配置采样过程，允许自定义生成参数。
- Comfy dtype: FLOAT
- Python dtype: float
- denoise
- 未知
- Comfy dtype: FLOAT
- Python dtype: unknown
- refiner_steps
- 未知
- Comfy dtype: INT
- Python dtype: unknown
- refiner_cfg
- 未知
- Comfy dtype: FLOAT
- Python dtype: unknown
- refiner_denoise
- 未知
- Comfy dtype: FLOAT
- Python dtype: unknown
- sampler_name
- 标识要使用的特定采样器算法，影响生成图像的风格和特征。
- Comfy dtype: COMBO[STRING]
- Python dtype: str
- scheduler
- 指定采样过程的调度算法，影响图像生成的进程和动态。
- Comfy dtype: COMBO[STRING]
- Python dtype: str
- image_output
- 决定生成图像的输出格式和处理方式，包括保存或显示选项。
- Comfy dtype: COMBO[STRING]
- Python dtype: str
- save_prefix
- 设置保存图像文件的前缀，组织和识别生成的图像。
- Comfy dtype: STRING
- Python dtype: str
Optional
- seed
- 未知
- Comfy dtype: INT
- Python dtype: unknown
- optional_model
- 未知
- Comfy dtype: MODEL
- Python dtype: unknown
- optional_positive
- 未知
- Comfy dtype: CONDITIONING
- Python dtype: unknown
- optional_negative
- 未知
- Comfy dtype: CONDITIONING
- Python dtype: unknown
- optional_latent
- 未知
- Comfy dtype: LATENT
- Python dtype: unknown
- optional_vae
- 未知
- Comfy dtype: VAE
- Python dtype: unknown
- optional_refiner_model
- 未知
- Comfy dtype: MODEL
- Python dtype: unknown
- optional_refiner_positive
- 未知
- Comfy dtype: CONDITIONING
- Python dtype: unknown
- optional_refiner_negative
- 未知
- Comfy dtype: CONDITIONING
- Python dtype: unknown
- optional_clip
- 未知
- Comfy dtype: CLIP
- Python dtype: unknown
- input_image_override
- 未知
- Comfy dtype: IMAGE
- Python dtype: unknown
- adv_xyPlot
- 未知
- Comfy dtype: ADV_XYPLOT
- Python dtype: unknown
Output types
- sdxl_pipe
- Comfy dtype: PIPE_LINE_SDXL
- 未知
- Python dtype: unknown
- pipe
- Comfy dtype: PIPE_LINE
- 未知
- Python dtype: unknown
- model
- Comfy dtype: MODEL
- 未知
- Python dtype: unknown
- positive
- Comfy dtype: CONDITIONING
- 未知
- Python dtype: unknown
- negative
- Comfy dtype: CONDITIONING
- 未知
- Python dtype: unknown
- refiner_model
- Comfy dtype: MODEL
- 未知
- Python dtype: unknown
- refiner_positive
- Comfy dtype: CONDITIONING
- 未知
- Python dtype: unknown
- refiner_negative
- Comfy dtype: CONDITIONING
- 未知
- Python dtype: unknown
- latent
- Comfy dtype: LATENT
- 未知
- Python dtype: unknown
- vae
- Comfy dtype: VAE
- 未知
- Python dtype: unknown
- clip
- Comfy dtype: CLIP
- 未知
- Python dtype: unknown
- image
- Comfy dtype: IMAGE
- 未知
- Python dtype: unknown
- seed
- Comfy dtype: INT
- 未知
- Python dtype: unknown
- ui
- 提供一个用户界面组件，显示图像生成过程的结果，包括生成的图像和可能的额外信息。
Usage tips
- Infra type:
GPU - Common nodes: unknown
Source code
class ttN_pipeKSamplerSDXL_v2:
version = '2.1.0'
upscale_methods = ["None",
"[latent] nearest-exact", "[latent] bilinear", "[latent] area", "[latent] bicubic", "[latent] lanczos", "[latent] bislerp",
"[hiresFix] nearest-exact", "[hiresFix] bilinear", "[hiresFix] area", "[hiresFix] bicubic", "[hiresFix] lanczos", "[hiresFix] bislerp"]
crop_methods = ["disabled", "center"]
def __init__(self):
pass
@classmethod
def INPUT_TYPES(cls):
return {"required":
{"sdxl_pipe": ("PIPE_LINE_SDXL",),
"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": (cls.upscale_methods, {"default": "None"}),
"upscale_model_name": (folder_paths.get_filename_list("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": (cls.crop_methods,),
"sampler_state": (["Sample", "Hold"], ),
"base_steps": ("INT", {"default": 20, "min": 1, "max": 10000}),
"cfg": ("FLOAT", {"default": 8.0, "min": 0.0, "max": 100.0}),
"denoise": ("FLOAT", {"default": 1.0, "min": 0.0, "max": 1.0, "step": 0.01}),
"refiner_steps": ("INT", {"default": 20, "min": 0, "max": 10000}),
"refiner_cfg": ("FLOAT", {"default": 8.0, "min": 0.0, "max": 100.0}),
"refiner_denoise": ("FLOAT", {"default": 1.0, "min": 0.0, "max": 1.0, "step": 0.01}),
"sampler_name": (comfy.samplers.KSampler.SAMPLERS,),
"scheduler": (comfy.samplers.KSampler.SCHEDULERS,),
"image_output": (["Hide", "Preview", "Save", "Hide/Save", "Disabled"],),
"save_prefix": ("STRING", {"default": "ComfyUI"})
},
"optional":
{"seed": ("INT", {"default": 0, "min": 0, "max": 0xffffffffffffffff}),
"optional_model": ("MODEL",),
"optional_positive": ("CONDITIONING",),
"optional_negative": ("CONDITIONING",),
"optional_latent": ("LATENT",),
"optional_vae": ("VAE",),
"optional_refiner_model": ("MODEL",),
"optional_refiner_positive": ("CONDITIONING",),
"optional_refiner_negative": ("CONDITIONING",),
"optional_latent": ("LATENT",),
"optional_clip": ("CLIP",),
"input_image_override": ("IMAGE",),
"adv_xyPlot": ("ADV_XYPLOT",),
},
"hidden":
{"prompt": "PROMPT", "extra_pnginfo": "EXTRA_PNGINFO", "my_unique_id": "UNIQUE_ID",
"embeddingsList": (folder_paths.get_filename_list("embeddings"),),
"lorasList": (folder_paths.get_filename_list("loras"),),
"ttNnodeVersion": ttN_pipeKSamplerSDXL_v2.version},
}
RETURN_TYPES = ("PIPE_LINE_SDXL", "PIPE_LINE", "MODEL", "CONDITIONING", "CONDITIONING", "MODEL", "CONDITIONING", "CONDITIONING", "LATENT", "VAE", "CLIP", "IMAGE", "INT",)
RETURN_NAMES = ("sdxl_pipe", "pipe","model", "positive", "negative" , "refiner_model", "refiner_positive", "refiner_negative", "latent", "vae", "clip", "image", "seed", )
OUTPUT_NODE = True
FUNCTION = "sample"
CATEGORY = "ttN/pipe"
def sample(self, sdxl_pipe,
lora_name, lora_strength,
sampler_state, base_steps, refiner_steps, cfg, denoise, refiner_cfg, refiner_denoise, sampler_name, scheduler, image_output, save_prefix,
optional_model=None, optional_positive=None, optional_negative=None, optional_latent=None, optional_vae=None, optional_clip=None, input_image_override=None, adv_xyPlot=None,
seed=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, force_full_denoise=False, disable_noise=False,
optional_refiner_model=None, optional_refiner_positive=None, optional_refiner_negative=None):
# Clean Loader Models from Global
loader.update_loaded_objects(prompt)
my_unique_id = int(my_unique_id)
ttN_save = ttNsave(my_unique_id, prompt, extra_pnginfo)
sdxl_model = optional_model if optional_model is not None else sdxl_pipe["model"]
sdxl_positive = optional_positive if optional_positive is not None else sdxl_pipe["positive"]
sdxl_negative = optional_negative if optional_negative is not None else sdxl_pipe["negative"]
sdxl_samples = optional_latent if optional_latent is not None else sdxl_pipe["samples"]
sdxl_images = input_image_override if input_image_override is not None else sdxl_pipe["images"]
sdxl_vae = optional_vae if optional_vae is not None else sdxl_pipe["vae"]
sdxl_clip = optional_clip if optional_clip is not None else sdxl_pipe["clip"]
sdxl_refiner_model = optional_refiner_model if optional_refiner_model is not None else sdxl_pipe["refiner_model"]
sdxl_refiner_positive = optional_refiner_positive if optional_refiner_positive is not None else sdxl_pipe["refiner_positive"]
#sdxl_refiner_positive = sdxl_positive if sdxl_refiner_positive is None else sdxl_refiner_positive
sdxl_refiner_negative = optional_refiner_negative if optional_refiner_negative is not None else sdxl_pipe["refiner_negative"]
#sdxl_refiner_negative = sdxl_negative if sdxl_refiner_negative is None else sdxl_refiner_negative
sdxl_refiner_clip = sdxl_pipe["refiner_clip"]
if seed in (None, 'undefined'):
sdxl_seed = sdxl_pipe["seed"]
else:
sdxl_seed = seed
def process_sample_state(sdxl_pipe, sdxl_model, sdxl_images, sdxl_clip, sdxl_samples, sdxl_vae, sdxl_seed, sdxl_positive, sdxl_negative, lora_name, lora_model_strength, lora_clip_strength,
sdxl_refiner_model, sdxl_refiner_positive, sdxl_refiner_negative, sdxl_refiner_clip,
upscale_model_name, upscale_method, factor, rescale, percent, width, height, longer_side, crop,
base_steps, refiner_steps, cfg, sampler_name, scheduler, denoise, refiner_denoise,
image_output, save_prefix, prompt, my_unique_id, preview_latent, force_full_denoise=force_full_denoise, disable_noise=disable_noise):
# Load Lora
if lora_name not in (None, "None"):
sdxl_model, sdxl_clip = loader.load_lora(lora_name, sdxl_model, sdxl_clip, lora_model_strength, lora_clip_strength)
total_steps = base_steps + refiner_steps
# Upscale samples if enabled
upscale_method = upscale_method.split(' ', 1)
if upscale_method[0] == "[latent]":
sdxl_samples = sampler.handle_upscale(sdxl_samples, upscale_method[1], factor, crop)
if upscale_method[0] == "[hiresFix]":
if (sdxl_images is None):
sdxl_images = sdxl_vae.decode(sdxl_samples["samples"])
hiresfix = ttN_modelScale()
sdxl_samples = hiresfix.upscale(upscale_model_name, sdxl_images, True if rescale != 'None' else False, upscale_method[1], rescale, percent, width, height, longer_side, crop, "return latent", None, True, sdxl_vae)
if (refiner_steps > 0) and (sdxl_refiner_model not in [None, "None"]):
# Base Sample
sdxl_samples = sampler.common_ksampler(sdxl_model, sdxl_seed, total_steps, cfg, sampler_name, scheduler, sdxl_positive, sdxl_negative, sdxl_samples,
denoise=denoise, preview_latent=preview_latent, start_step=0, last_step=base_steps, force_full_denoise=force_full_denoise, disable_noise=disable_noise)
# Refiner Sample
sdxl_samples = sampler.common_ksampler(sdxl_refiner_model, sdxl_seed, total_steps, refiner_cfg, sampler_name, scheduler, sdxl_refiner_positive, sdxl_refiner_negative, sdxl_samples,
denoise=refiner_denoise, preview_latent=preview_latent, start_step=base_steps, last_step=10000, force_full_denoise=True, disable_noise=True)
else:
sdxl_samples = sampler.common_ksampler(sdxl_model, sdxl_seed, base_steps, cfg, sampler_name, scheduler, sdxl_positive, sdxl_negative, sdxl_samples,
denoise=denoise, preview_latent=preview_latent, start_step=0, last_step=base_steps, force_full_denoise=True, disable_noise=disable_noise)
results = list()
if (image_output != "Disabled"):
latent = sdxl_samples["samples"]
sdxl_images = sdxl_vae.decode(latent)
results = ttN_save.images(sdxl_images, save_prefix, image_output)
sampler.update_value_by_id("results", my_unique_id, results)
# Clean loaded_objects
loader.update_loaded_objects(prompt)
new_sdxl_pipe = {
"model": sdxl_model,
"positive": sdxl_positive,
"negative": sdxl_negative,
"vae": sdxl_vae,
"clip": sdxl_clip,
"refiner_model": sdxl_refiner_model,
"refiner_positive": sdxl_refiner_positive,
"refiner_negative": sdxl_refiner_negative,
"refiner_clip": sdxl_refiner_clip,
"samples": sdxl_samples,
"images": sdxl_images,
"seed": sdxl_seed,
"loader_settings": sdxl_pipe["loader_settings"],
}
pipe = {"model": sdxl_model,
"positive": sdxl_positive,
"negative": sdxl_negative,
"vae": sdxl_vae,
"clip": sdxl_clip,
"samples": sdxl_samples,
"images": sdxl_images,
"seed": sdxl_seed,
"loader_settings": sdxl_pipe["loader_settings"],
}
sampler.update_value_by_id("pipe_line", my_unique_id, new_sdxl_pipe)
del sdxl_pipe
if image_output in ("Hide", "Hide/Save", "Disabled"):
return sampler.get_output_sdxl_v2(new_sdxl_pipe, pipe)
return {"ui": {"images": results},
"result": sampler.get_output_sdxl_v2(new_sdxl_pipe, pipe)}
def process_xyPlot(sdxl_pipe, sdxl_model, sdxl_clip, sdxl_samples, sdxl_vae, sdxl_seed, sdxl_positive, sdxl_negative, lora_name, lora_model_strength, lora_clip_strength,
base_steps, refiner_steps, cfg, sampler_name, scheduler, denoise,
image_output, save_prefix, prompt, extra_pnginfo, my_unique_id, preview_latent, adv_xyPlot):
random.seed(seed)
plotter = ttNadv_xyPlot(adv_xyPlot, my_unique_id, prompt, extra_pnginfo, save_prefix, image_output)
samples, images = plotter.xy_plot_process()
if samples is None and images is None:
return process_sample_state(sdxl_pipe, sdxl_model, sdxl_images, sdxl_clip, sdxl_samples, sdxl_vae, sdxl_seed, sdxl_positive, sdxl_negative, lora_name, lora_model_strength, lora_clip_strength,
sdxl_refiner_model, sdxl_refiner_positive, sdxl_refiner_negative, sdxl_refiner_clip,
upscale_model_name, upscale_method, factor, rescale, percent, width, height, longer_side, crop,
base_steps, refiner_steps, cfg, sampler_name, scheduler, denoise, refiner_denoise,
image_output, save_prefix, prompt, my_unique_id, preview_latent, force_full_denoise=force_full_denoise, disable_noise=disable_noise)
results = ttN_save.images(images[0], save_prefix, image_output)
sampler.update_value_by_id("results", my_unique_id, results)
# Clean loaded_objects
loader.update_loaded_objects(prompt)
new_sdxl_pipe = {
"model": sdxl_model,
"positive": sdxl_positive,
"negative": sdxl_negative,
"vae": sdxl_vae,
"clip": sdxl_clip,
"refiner_model": sdxl_refiner_model,
"refiner_positive": sdxl_refiner_positive,
"refiner_negative": sdxl_refiner_negative,
"refiner_clip": sdxl_refiner_clip,
"samples": samples,
"images": images[1],
"seed": sdxl_seed,
"loader_settings": sdxl_pipe["loader_settings"],
}
pipe = {"model": sdxl_model,
"positive": sdxl_positive,
"negative": sdxl_negative,
"vae": sdxl_vae,
"clip": sdxl_clip,
"samples": samples,
"images": images[1],
"seed": sdxl_seed,
"loader_settings": sdxl_pipe["loader_settings"],
}
sampler.update_value_by_id("pipe_line", my_unique_id, new_sdxl_pipe)
del sdxl_pipe
if image_output in ("Hide", "Hide/Save", "Disabled"):
return sampler.get_output_sdxl_v2(new_sdxl_pipe, pipe)
return {"ui": {"images": results},
"result": sampler.get_output_sdxl_v2(new_sdxl_pipe, pipe)}
preview_latent = True
if image_output in ("Hide", "Hide/Save", "Disabled"):
preview_latent = False
if sampler_state == "Sample" and adv_xyPlot is None:
return process_sample_state(sdxl_pipe, sdxl_model, sdxl_images, sdxl_clip, sdxl_samples, sdxl_vae, sdxl_seed, sdxl_positive, sdxl_negative,
lora_name, lora_strength, lora_strength,
sdxl_refiner_model, sdxl_refiner_positive, sdxl_refiner_negative, sdxl_refiner_clip,
upscale_model_name, upscale_method, factor, rescale, percent, width, height, longer_side, crop,
base_steps, refiner_steps, cfg, sampler_name, scheduler, denoise, refiner_denoise, image_output, save_prefix, prompt, my_unique_id, preview_latent)
elif sampler_state == "Sample" and adv_xyPlot is not None:
return process_xyPlot(sdxl_pipe, sdxl_model, sdxl_clip, sdxl_samples, sdxl_vae, sdxl_seed, sdxl_positive, sdxl_negative, lora_name, lora_strength, lora_strength,
base_steps, refiner_steps, cfg, sampler_name, scheduler, denoise,
image_output, save_prefix, prompt, extra_pnginfo, my_unique_id, preview_latent, adv_xyPlot)
elif sampler_state == "Hold":
return sampler.process_hold_state(sdxl_pipe, image_output, my_unique_id, sdxl=True)