ComfyUI Node
Inputs Diffusers Refine Basic
A lightweight diffusion-model setup node that loads a standalone UNet with explicit CLIP and VAE files, prepares common sampler settings, creates an empty latent batch, and optionally provides a separate refinement diffusion model. Right-click → "Collapse Connections" or double-click to hide output labels and converge connection lines.
Inputs Diffusers Refine Basic
- MODEL
- CLIP
- VAE
- MODEL REFINE
- model name
- model name refine
- sampler
- scheduler
- cfg
- steps
- latent
- width
- height
- seed
- batch size
- speed lora ratio
◄diffusion_name▾►
◄refine_diffusion_nameNone►
◄clip_name▾►
◄speed_lorafalse►
◄speed_lora_modelNone►
◄speed_lora_strength1.00►
◄speed_lora_full_bothfalse►
◄speed_lora_ratio1.00►
◄clip_typeAce►
◄vae_name▾►
◄sampler▾►
◄scheduler▾►
◄cfg2.0►
◄steps10►
◄width720►
◄height720►
◄aspect_ratio▾►
◄aspect_modeOriginal►
◄batch_size1►
◄seed0►
◄megapixels1.0►
◄multiple8►
CategoryAUN Nodes/Loaders+Inputs
Inputs (22)
| Name | Type | Default | Description |
|---|---|---|---|
| diffusion_name | COMBO | Primary diffusion-model file (UNet only). Matches the comfy-core diffusion model list. | |
| refine_diffusion_name | COMBO | None | Optional refinement diffusion-model file. Select 'None' to reuse the main model. |
| clip_name | COMBO | CLIP file shared by the main and refine diffusion models. | |
| speed_lora | BOOLEAN | false | Enable SpeedLoRA when loading models. |
| speed_lora_model | COMBO | None | SpeedLoRA file to apply after loading the models. |
| speed_lora_strength | FLOAT | 1.000–3 | Strength multiplier when applying the SpeedLoRA weights. |
| speed_lora_full_both | BOOLEAN | false | Apply the full SpeedLoRA strength to both the main and refine models. |
| speed_lora_ratio | FLOAT | 1.000–1 | Share of the SpeedLoRA strength applied to the main model. The refine model receives the remaining share. |
| clip_type | COMBO | Ace | Clip architecture to use when loading a diffusion model. |
| vae_name | COMBO | VAE checkpoint shared by the main and refine diffusion models. | |
| sampler | COMBO | Sampling algorithm. | |
| scheduler | COMBO | Noise schedule. | |
| cfg | FLOAT | 2.0-2–100 | Classifier-Free Guidance scale. |
| steps | INT | 101–10000 | Sampling steps. |
| width | INT | 72064–8192 | Image width when aspect_ratio='custom'. |
| height | INT | 72064–8192 | Image height when aspect_ratio='custom'. |
| aspect_ratio | COMBO | Preset aspect ratio that overrides width/height. | |
| aspect_mode | COMBO | Original | Random swaps dimensions 50% of the time, Swap always flips width/height. |
| batch_size | INT | 11–64 | Latent batch size. |
| seed | INT | 00–18446744073709550000 | Base seed. |
| megapixels | FLOAT | 1.00.1–16 | Target total megapixels used when a ratio is selected. |
| multiple | INT | 88–128 | Nearest multiple to round computed resolution to. Used with ratio. |
Outputs (16)
| Name | Type | Description |
|---|---|---|
| MODEL | MODEL | — |
| CLIP | CLIP | — |
| VAE | VAE | — |
| MODEL REFINE | MODEL | — |
| model name | STRING | — |
| model name refine | STRING | — |
| sampler | * | — |
| scheduler | * | — |
| cfg | FLOAT | — |
| steps | INT | — |
| latent | LATENT | — |
| width | INT | — |
| height | INT | — |
| seed | INT | — |
| batch size | INT | — |
| speed lora ratio | FLOAT | — |