4.8 KiB
4.8 KiB
Documentation
- Class name: MikeySampler
- Category: Mikey/Sampling
- Output node: False
- Repo Ref: https://github.com/bash-j/mikey_nodes
MikeySampler节点旨在从给定的基础模型和细化模型生成高质量的样本，同时利用VAE进行潜在空间操作。它执行复杂的采样技术，并结合条件输入来优化输出。该节点还根据指定的方法和强度对生成的图像进行放大，确保最终输出满足所需的分辨率和质量标准。
Input types
Required
- seed
- 种子参数对于采样过程的可重复性至关重要，确保在不同运行中生成的样本是一致的。它在采样算法的随机性控制中起着重要作用。
- Comfy dtype: INT
- Python dtype: int
- base_model
- 基础模型是用于潜在空间初始采样的基础神经网络。它是一个关键组件，决定了生成样本的质量和多样性。
- Comfy dtype: MODEL
- Python dtype: torch.nn.Module
- refiner_model
- 细化模型用于进一步提高基础模型生成的样本质量。它细化潜在空间表示，以实现最终输出的更高保真度。
- Comfy dtype: MODEL
- Python dtype: torch.nn.Module
- samples
- 样本参数表示要从基础模型中采样的潜在向量的数量。它直接影响生成输出的多样性。
- Comfy dtype: LATENT
- Python dtype: torch.Tensor
Optional
- upscale_by
- upscale_by参数确定生成图像的缩放因子。它是控制最终输出分辨率的重要因素。
- Comfy dtype: FLOAT
- Python dtype: float
- hires_strength
- hires_strength参数调整高分辨率放大过程的强度。它影响放大图像的清晰度和细节水平。
- Comfy dtype: FLOAT
- Python dtype: float
Output types
- latent
- latent参数表示采样过程中得到的编码潜在向量。它很重要，因为它作为模型和最终放大图像之间的中间表示。
- Comfy dtype: LATENT
- Python dtype: torch.Tensor
Usage tips
- Infra type: GPU
Source code
class MikeySampler:
@classmethod
def INPUT_TYPES(s):
return {'required': {'base_model': ('MODEL',), 'refiner_model': ('MODEL',), 'samples': ('LATENT',), 'vae': ('VAE',), 'positive_cond_base': ('CONDITIONING',), 'negative_cond_base': ('CONDITIONING',), 'positive_cond_refiner': ('CONDITIONING',), 'negative_cond_refiner': ('CONDITIONING',), 'model_name': (folder_paths.get_filename_list('upscale_models'),), 'seed': ('INT', {'default': 0, 'min': 0, 'max': 18446744073709551615}), 'upscale_by': ('FLOAT', {'default': 1.0, 'min': 0.0, 'max': 10.0, 'step': 0.1}), 'hires_strength': ('FLOAT', {'default': 1.0, 'min': 0.0, 'max': 2.0, 'step': 0.1})}}
RETURN_TYPES = ('LATENT',)
FUNCTION = 'run'
CATEGORY = 'Mikey/Sampling'
def adjust_start_step(self, image_complexity, hires_strength=1.0):
image_complexity /= 24
if image_complexity > 1:
image_complexity = 1
image_complexity = min([0.55, image_complexity]) * hires_strength
return min([16, 16 - int(round(image_complexity * 16, 0))])
def run(self, seed, base_model, refiner_model, vae, samples, positive_cond_base, negative_cond_base, positive_cond_refiner, negative_cond_refiner, model_name, upscale_by=1.0, hires_strength=1.0, upscale_method='normal'):
image_scaler = ImageScale()
vaeencoder = VAEEncode()
vaedecoder = VAEDecode()
uml = UpscaleModelLoader()
upscale_model = uml.load_model(model_name)[0]
iuwm = ImageUpscaleWithModel()
sample1 = common_ksampler(base_model, seed, 25, 6.5, 'dpmpp_2s_ancestral', 'simple', positive_cond_base, negative_cond_base, samples, start_step=0, last_step=18, force_full_denoise=False)[0]
sample2 = common_ksampler(refiner_model, seed, 30, 3.5, 'dpmpp_2m', 'simple', positive_cond_refiner, negative_cond_refiner, sample1, disable_noise=True, start_step=21, force_full_denoise=True)
if upscale_by == 0:
return sample2
else:
sample2 = sample2[0]
pixels = vaedecoder.decode(vae, sample2)[0]
(org_width, org_height) = (pixels.shape[2], pixels.shape[1])
img = iuwm.upscale(upscale_model, image=pixels)[0]
(upscaled_width, upscaled_height) = (int(org_width * upscale_by // 8 * 8), int(org_height * upscale_by // 8 * 8))
img = image_scaler.upscale(img, 'nearest-exact', upscaled_width, upscaled_height, 'center')[0]
if hires_strength == 0:
return (vaeencoder.encode(vae, img)[0],)
image_complexity = calculate_image_complexity(img)
start_step = self.adjust_start_step(image_complexity, hires_strength)
latent = vaeencoder.encode(vae, img)[0]
out = common_ksampler(base_model, seed, 16, 9.5, 'dpmpp_2m_sde', 'karras', positive_cond_base, negative_cond_base, latent, start_step=start_step, force_full_denoise=True)
return out