Mikey Sampler Base Only

Documentation

• Class name: Mikey Sampler Base Only
• Category: Mikey/Sampling
• Output node: False

The Mikey Sampler Base Only node is designed for basic sampling tasks within a specialized framework, focusing on generating outputs based on a simplified set of parameters. It abstracts the complexity of underlying sampling algorithms to provide a straightforward interface for generating samples.

Input types

Required

• base_model
  • Specifies the base model used for sampling, serving as the foundation for generating outputs.
  • Comfy dtype: MODEL
  • Python dtype: str
• samples
  • Represents the latent samples that are either inputted to or generated by the node, depending on the context.
  • Comfy dtype: LATENT
  • Python dtype: str
• positive_cond_base
  • Defines the positive conditioning for the base model, guiding the generation towards desired outcomes.
  • Comfy dtype: CONDITIONING
  • Python dtype: str
• negative_cond_base
  • Specifies the negative conditioning for the base model, aiming to steer the generation away from certain themes or elements.
  • Comfy dtype: CONDITIONING
  • Python dtype: str
• vae
  • The variational autoencoder used in conjunction with the base model to enhance the sampling process.
  • Comfy dtype: VAE
  • Python dtype: str
• model_name
  • The name of the model used for upsampling, providing a means to refine the generated samples.
  • Comfy dtype: COMBO[STRING]
  • Python dtype: str
• seed
  • A seed value for random number generation, ensuring reproducibility of the samples.
  • Comfy dtype: INT
  • Python dtype: int
• upscale_by
  • A factor by which the samples are upscaled, affecting the resolution and detail of the output.
  • Comfy dtype: FLOAT
  • Python dtype: float
• hires_strength
  • Determines the strength of high-resolution features in the upscaled samples, allowing for finer control over the output quality.
  • Comfy dtype: FLOAT
  • Python dtype: float
• smooth_step
  • An optional parameter that can smooth the transition between steps in the sampling process, potentially enhancing the output quality.
  • Comfy dtype: INT
  • Python dtype: int

Output types

• latent
  • Comfy dtype: LATENT
  • The output is a latent representation of the generated sample, encapsulating the complex features and characteristics derived from the input prompts.
  • Python dtype: str

Usage tips

• Infra type: CPU
• Common nodes: unknown

Source code

class MikeySamplerBaseOnly:
    @classmethod
    def INPUT_TYPES(s):
        return {"required": {"base_model": ("MODEL",), "samples": ("LATENT",),
                             "positive_cond_base": ("CONDITIONING",), "negative_cond_base": ("CONDITIONING",),
                             "vae": ("VAE",),
                             "model_name": (folder_paths.get_filename_list("upscale_models"), ),
                             "seed": ("INT", {"default": 0, "min": 0, "max": 0xffffffffffffffff}),
                             "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}),
                             'smooth_step': ("INT", {"default": 0, "min": -1, "max": 100})}}

    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([31, 31 - int(round(image_complexity * 31,0))])

    def run(self, seed, base_model, vae, samples, positive_cond_base, negative_cond_base,
            model_name, upscale_by=1.0, hires_strength=1.0, upscale_method='normal', smooth_step=0):
        image_scaler = ImageScale()
        vaeencoder = VAEEncode()
        vaedecoder = VAEDecode()
        uml = UpscaleModelLoader()
        upscale_model = uml.load_model(model_name)[0]
        iuwm = ImageUpscaleWithModel()
        # common_ksampler(model, seed, steps, cfg, sampler_name, scheduler, positive, negative, latent, denoise=1.0,
        # disable_noise=False, start_step=None, last_step=None, force_full_denoise=False)
        # step 1 run base model low cfg
        sample1 = common_ksampler(base_model, seed, 30, 4, 'dpmpp_2m_sde_gpu', 'karras', positive_cond_base, negative_cond_base, samples,
                                  start_step=0, last_step=14, force_full_denoise=False)[0]
        # step 2 run base model high cfg
        sample2 = common_ksampler(base_model, seed+1, 31 + smooth_step, 6, 'dpmpp_2m_sde_gpu', 'karras', positive_cond_base, negative_cond_base, sample1,
                                  disable_noise=True, start_step=15, force_full_denoise=True)
        if upscale_by == 0:
            return sample2
        else:
            sample2 = sample2[0]
        # step 3 upscale
        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],)
        # Adjust start_step based on complexity
        image_complexity = calculate_image_complexity(img)
        #print('Image Complexity:', image_complexity)
        start_step = self.adjust_start_step(image_complexity, hires_strength)
        # encode image
        latent = vaeencoder.encode(vae, img)[0]
        # step 3 run base model
        out = common_ksampler(base_model, seed, 31, 6, 'dpmpp_2m_sde_gpu', 'karras', positive_cond_base, negative_cond_base, latent,
                                start_step=start_step, force_full_denoise=True)
        return out