Load Diffusion Model Hub Pro

Central parameter hub that loads a diffusion/UNET model and routes MODEL, steps, refiner steps, CFG, sampler algorithm, scheduler curve, denoise values, positive/negative prompts, and metadata across workflows.

By MarwanDSAI·Created 15 days ago·Updated 4 days ago· 1

MODEL
steps
step_refiner
cfg
unet_name
sampler_name
scheduler
denoise
positive_prompt
negative_prompt
metadata

◄unet_nameNone►

◄steps20►

◄step_refiner10►

◄cfg7.0►

◄sampler_name▾►

◄scheduler▾►

◄denoise1.00►

◄positive_prompt►

◄negative_prompt►

◄metadata►

CategoryMDSNodes/sampling

Inputs (10)

Name	Type	Default	Description
unet_name	COMBO	None	Select the diffusion / UNET model file from your models/diffusion_models or models/unet folder.
steps	INT	201–10000	The total number of sampling/denoising steps for base generation.
step_refiner	INT	100–10000	Target step count for refiner passes or the step transition threshold for multi-pass pipelines.
cfg	FLOAT	7.00–100	Classifier-Free Guidance (CFG) scale. Controls how strictly the model adheres to your prompt.
sampler_name	COMBO		The mathematical sampling algorithm used to generate or denoise the image (e.g., euler, dpmpp_2m).
scheduler	COMBO		The noise scheduling rate/curve across the steps (e.g., normal, karras, sgm_uniform, simple).
denoise	FLOAT	1.000–1	Denoise strength. Set to 1.0 for initial txt2img generation, or 0.20-0.60 for img2img / upscaling.
positive_prompt	STRING		Enter positive prompt text. Expanding the node will enlarge this text box.
negative_prompt	STRING		Enter negative prompt text. Expanding the node will enlarge this text box.
metadata	STRING		Single-line metadata text or workflow tags to pass downstream.

Outputs (11)

Name	Type	Description
MODEL	MODEL	The loaded diffusion model.
steps	INT	Base generation step count (INT).
step_refiner	INT	Refiner step count or threshold (INT).
cfg	FLOAT	Classifier-Free Guidance scale (FLOAT).
unet_name	*	The selected diffusion/UNET model filename (Universal Wildcard *).
sampler_name	euler,euler_cfg_pp,euler_ancestral,euler_ancestral_cfg_pp,heun,heunpp2,exp_heun_2_x0,exp_heun_2_x0_sde,dpm_2,dpm_2_ancestral,lms,dpm_fast,dpm_adaptive,dpmpp_2s_ancestral,dpmpp_2s_ancestral_cfg_pp,dpmpp_sde,dpmpp_sde_gpu,dpmpp_2m,dpmpp_2m_cfg_pp,dpmpp_2m_sde,dpmpp_2m_sde_gpu,dpmpp_2m_sde_heun,dpmpp_2m_sde_heun_gpu,dpmpp_3m_sde,dpmpp_3m_sde_gpu,ddpm,lcm,ipndm,ipndm_v,deis,res_multistep,res_multistep_cfg_pp,res_multistep_ancestral,res_multistep_ancestral_cfg_pp,gradient_estimation,gradient_estimation_cfg_pp,er_sde,seeds_2,seeds_3,sa_solver,sa_solver_pece,ddim,uni_pc,uni_pc_bh2	Sampler algorithm name (COMBO slot for KSampler / SamplerSelect).
scheduler	simple,sgm_uniform,karras,exponential,ddim_uniform,beta,normal,linear_quadratic,kl_optimal	Scheduler curve type (COMBO slot for KSampler / BasicScheduler).
denoise	FLOAT	Denoise strength multiplier (FLOAT).
positive_prompt	STRING	Positive prompt string (connect to CLIP Text Encode).
negative_prompt	STRING	Negative prompt string (connect to CLIP Text Encode).
metadata	STRING	Passthrough metadata string.