RegionalSamplerAdvanced¶
Documentation¶
- Class name:
RegionalSamplerAdvanced - Category:
ImpactPack/Regional - Output node:
False
The RegionalSamplerAdvanced node is designed for advanced sampling operations within specific regions of latent images. It utilizes a base sampler and additional samplers to apply complex sampling strategies, enabling precise control over the sampling process in targeted areas.
Input types¶
Required¶
add_noise- Determines whether noise is added to the sampling process, affecting the texture and details of the sampled output.
- Comfy dtype:
BOOLEAN - Python dtype:
bool
noise_seed- Sets the seed for noise generation, ensuring reproducibility of the noise patterns in the sampling process.
- Comfy dtype:
INT - Python dtype:
int
steps- Specifies the number of steps to perform in the sampling process, influencing the depth and detail of the sampling.
- Comfy dtype:
INT - Python dtype:
int
start_at_step- Defines the starting step for the sampling process, allowing for partial sampling starting from a specific point.
- Comfy dtype:
INT - Python dtype:
int
end_at_step- Determines the ending step for the sampling process, enabling partial sampling up to a specific point.
- Comfy dtype:
INT - Python dtype:
int
overlap_factor- Controls the overlap between sampled regions, affecting the blending and transition between sampled areas.
- Comfy dtype:
INT - Python dtype:
int
restore_latent- Indicates whether the original latent image is restored after sampling, preserving the original content outside the sampled regions.
- Comfy dtype:
BOOLEAN - Python dtype:
bool
return_with_leftover_noise- Determines whether the sampled output includes leftover noise, adding an additional layer of texture to the output.
- Comfy dtype:
BOOLEAN - Python dtype:
bool
latent_image- The latent image to be sampled, serving as the base for the sampling operations.
- Comfy dtype:
LATENT - Python dtype:
torch.Tensor
base_sampler- The base sampler used for the initial sampling process, setting the foundation for further sampling operations.
- Comfy dtype:
KSAMPLER_ADVANCED - Python dtype:
str
regional_prompts- Specifies the regional prompts used to guide the sampling process in specific areas, enhancing the relevance and accuracy of the sampled output.
- Comfy dtype:
REGIONAL_PROMPTS - Python dtype:
List[Dict]
additional_mode- Defines the mode for additional sampling operations, allowing for customization of the sampling strategy.
- Comfy dtype:
COMBO[STRING] - Python dtype:
str
additional_sampler- Specifies the additional sampler used for enhanced sampling operations, providing flexibility in the sampling approach.
- Comfy dtype:
COMBO[STRING] - Python dtype:
str
additional_sigma_ratio- Sets the sigma ratio for additional sampling operations, influencing the variance and intensity of the sampling.
- Comfy dtype:
FLOAT - Python dtype:
float
Output types¶
latent- Comfy dtype:
LATENT - The latent image after the advanced regional sampling process, reflecting the targeted modifications and enhancements.
- Python dtype:
torch.Tensor
- Comfy dtype:
Usage tips¶
- Infra type:
GPU - Common nodes: unknown
Source code¶
class RegionalSamplerAdvanced:
@classmethod
def INPUT_TYPES(s):
return {"required": {
"add_noise": ("BOOLEAN", {"default": True, "label_on": "enabled", "label_off": "disabled"}),
"noise_seed": ("INT", {"default": 0, "min": 0, "max": 0xffffffffffffffff}),
"steps": ("INT", {"default": 20, "min": 1, "max": 10000}),
"start_at_step": ("INT", {"default": 0, "min": 0, "max": 10000}),
"end_at_step": ("INT", {"default": 10000, "min": 0, "max": 10000}),
"overlap_factor": ("INT", {"default": 10, "min": 0, "max": 10000}),
"restore_latent": ("BOOLEAN", {"default": True, "label_on": "enabled", "label_off": "disabled"}),
"return_with_leftover_noise": ("BOOLEAN", {"default": False, "label_on": "enabled", "label_off": "disabled"}),
"latent_image": ("LATENT", ),
"base_sampler": ("KSAMPLER_ADVANCED", ),
"regional_prompts": ("REGIONAL_PROMPTS", ),
"additional_mode": (["DISABLE", "ratio additional", "ratio between"], {"default": "ratio between"}),
"additional_sampler": (["AUTO", "euler", "heun", "heunpp2", "dpm_2", "dpm_fast", "dpmpp_2m", "ddpm"],),
"additional_sigma_ratio": ("FLOAT", {"default": 0.3, "min": 0.0, "max": 1.0, "step": 0.01}),
},
"hidden": {"unique_id": "UNIQUE_ID"},
}
RETURN_TYPES = ("LATENT", )
FUNCTION = "doit"
CATEGORY = "ImpactPack/Regional"
def doit(self, add_noise, noise_seed, steps, start_at_step, end_at_step, overlap_factor, restore_latent, return_with_leftover_noise, latent_image, base_sampler, regional_prompts,
additional_mode, additional_sampler, additional_sigma_ratio, unique_id):
if restore_latent:
latent_compositor = nodes.NODE_CLASS_MAPPINGS['LatentCompositeMasked']()
else:
latent_compositor = None
masks = [regional_prompt.mask.numpy() for regional_prompt in regional_prompts]
masks = [np.ceil(mask).astype(np.int32) for mask in masks]
combined_mask = torch.from_numpy(np.bitwise_or.reduce(masks))
inv_mask = torch.where(combined_mask == 0, torch.tensor(1.0), torch.tensor(0.0))
region_len = len(regional_prompts)
end_at_step = min(steps, end_at_step)
total = (end_at_step - start_at_step) * region_len
new_latent_image = latent_image.copy()
base_latent_image = None
region_masks = {}
for i in range(start_at_step, end_at_step-1):
core.update_node_status(unique_id, f"{start_at_step+i}/{end_at_step} steps | ", ((i-start_at_step)*region_len)/total)
cur_add_noise = True if i == start_at_step and add_noise else False
new_latent_image['noise_mask'] = inv_mask
new_latent_image = base_sampler.sample_advanced(cur_add_noise, noise_seed, steps, new_latent_image, i, i + 1, True,
recovery_mode=additional_mode, recovery_sampler=additional_sampler, recovery_sigma_ratio=additional_sigma_ratio)
if restore_latent:
del new_latent_image['noise_mask']
base_latent_image = new_latent_image.copy()
j = 1
for regional_prompt in regional_prompts:
if restore_latent:
new_latent_image = base_latent_image.copy()
core.update_node_status(unique_id, f"{start_at_step+i}/{end_at_step} steps | {j}/{region_len}", ((i-start_at_step)*region_len + j)/total)
if j not in region_masks:
region_mask = regional_prompt.get_mask_erosion(overlap_factor).squeeze(0).squeeze(0)
region_masks[j] = region_mask
else:
region_mask = region_masks[j]
new_latent_image['noise_mask'] = region_mask
new_latent_image = regional_prompt.sampler.sample_advanced(False, noise_seed, steps, new_latent_image, i, i + 1, True,
recovery_mode=additional_mode, recovery_sampler=additional_sampler, recovery_sigma_ratio=additional_sigma_ratio)
if restore_latent:
del new_latent_image['noise_mask']
base_latent_image = latent_compositor.composite(base_latent_image, new_latent_image, 0, 0, False, region_mask)[0]
new_latent_image = base_latent_image
j += 1
# finalize
core.update_node_status(unique_id, f"finalize")
if base_latent_image is not None:
new_latent_image = base_latent_image
else:
base_latent_image = new_latent_image
new_latent_image['noise_mask'] = inv_mask
new_latent_image = base_sampler.sample_advanced(False, noise_seed, steps, new_latent_image, end_at_step-1, end_at_step, return_with_leftover_noise,
recovery_mode=additional_mode, recovery_sampler=additional_sampler, recovery_sigma_ratio=additional_sigma_ratio)
core.update_node_status(unique_id, f"{end_at_step}/{end_at_step} steps", total)
core.update_node_status(unique_id, "", None)
if restore_latent:
new_latent_image = base_latent_image
if 'noise_mask' in new_latent_image:
del new_latent_image['noise_mask']
return (new_latent_image, )