EasyUse/Loadersgenerated

EasyLoader (Comfy)(comfyLoader)

The EasyLoader (Comfy) node is an all-in-one loader from the ComfyUI-Easy-Use package.

comfy Loader

optional_lora_stack
optional_controlnet_stack
pipe
model
vae
ckpt_name
COMBO
vae_name
COMBO
clip_skip
-2
lora_name
COMBO
lora_model_strength
1.00
lora_clip_strength
1.00
resolution
512 x 512
empty_latent_width
INT
empty_latent_height
INT
positive
STRING
negative
STRING
batch_size
1
Easy Use

Description

The EasyLoader (Comfy) node is an all-in-one loader from the ComfyUI-Easy-Use package. It loads a checkpoint model, a VAE, a LoRA, applies CLIP skip, creates an empty latent image, and prepares positive and negative conditioning in one step. It is designed to simplify common setup tasks when building a generation workflow.

Inputs

NameTypeDefaultDescription
ckpt_nameCOMBOCheckpoint to load from the available model list.
vae_nameCOMBOVAE to load. Select a dedicated VAE file, or a VAE baked into the checkpoint if available.
clip_skipINT-2Number of final CLIP layers to skip. Negative values skip from the end; -1 removes one layer, -2 removes two, and so on. Range is -24 to 0.
lora_nameCOMBOLoRA to apply to the loaded model.
lora_model_strengthFLOAT1.0Strength applied to the LoRA's effect on the model. Range is -10.0 to 10.0 with a step of 0.01.
lora_clip_strengthFLOAT1.0Strength applied to the LoRA's effect on the CLIP model. Range is -10.0 to 10.0 with a step of 0.01.
resolutionCOMBO512 x 512Preset resolution shortcut. Selecting one of the common presets sets the empty latent dimensions to the corresponding width and height.
empty_latent_widthINTWidth of the empty latent image. This is the actual width used for generation when a custom size is needed.
empty_latent_heightINTHeight of the empty latent image. This is the actual height used for generation when a custom size is needed.
positiveSTRINGPositive prompt text used for the positive conditioning.
negativeSTRINGNegative prompt text used for the negative conditioning.
batch_sizeINT1Number of identical latent images to generate. Range is 1 to 4096.
optional_lora_stackLORA_STACKOptional LoRA stack input for applying additional LoRAs on top of the primary lora_name.
optional_controlnet_stackCONTROL_NET_STACKOptional ControlNet stack input for applying one or more ControlNets.

Outputs

TypeDescription
PIPE_LINEA pipeline object containing the loaded model, VAE, latent, positive and negative conditioning, and other common values. Use this with Easy-Use sampler or pipeline nodes.
MODELThe loaded model, after checkpoint loading and LoRA application.
VAEThe loaded VAE.

Usage Notes

  • The PIPE_LINE output is the most convenient output to pass into downstream Easy-Use nodes, since it bundles the model, VAE, latent, and conditioning together.
  • Use the MODEL and VAE outputs when connecting directly to standard ComfyUI nodes that expect separate model and VAE inputs.
  • clip_skip uses negative numbers: -1 skips the last CLIP layer, -2 skips the last two, and so on. A value of 0 disables CLIP skip.
  • If a resolution preset is chosen, it will replace the manual empty_latent_width and empty_latent_height values. For custom sizes, set the width and height directly.
  • The optional_lora_stack and optional_controlnet_stack inputs are optional. When left unconnected, the node runs using only the main LoRA and no ControlNet.
  • batch_size greater than 1 creates multiple copies of the empty latent, which can be useful for generating several images with the same conditioning in one pass.

Comments

Sign in with GitHub to join the discussion.

Loading comments…