ComfyUI Node: LoRA Optimizer (Legacy)

Authored by ethanfel

Created

Updated

135 stars

Run ComfyUI workflows without the setup

No installs, no CUDA version roulette, no GPU sitting idle on your bill. Bring a workflow and run it in the browser.

Category

LoRA Optimizer

Inputs

model MODEL
lora_stack LORA_STACK
output_strength FLOAT
clip CLIP
clip_strength_multiplier FLOAT
auto_strength
  • enabled
  • disabled
auto_strength_floor FLOAT
free_vram_between_passes
  • disabled
  • enabled
vram_budget FLOAT
optimization_mode
  • per_prefix
  • global
  • additive
cache_patches
  • enabled
  • disabled
patch_compression
  • smart
  • aggressive
  • disabled
svd_device
  • gpu
  • cpu
normalize_keys
  • disabled
  • enabled
sparsification
  • disabled
  • dare
  • della
  • dare_conflict
  • della_conflict
sparsification_density FLOAT
dare_dampening FLOAT
merge_refinement
  • none
  • refine
  • full
strategy_set
  • full
  • no_slerp
  • basic
architecture_preset
  • auto
  • sd_unet
  • dit
  • acestep_dit
  • llm
merge_strategy_override STRING
settings_source
  • manual
  • from_autotuner
  • from_tuner_data
tuner_data TUNER_DATA
decision_smoothing FLOAT
smooth_slerp_gate BOOLEAN

Outputs

MODEL

CLIP

STRING

TUNER_DATA

LORA_DATA

Extension: LoRA Optimizer

Auto-optimizer node for combining multiple LoRAs via diff-based merging with TIES conflict resolution

Authored by ethanfel

Looking for a different node?

Run ComfyUI workflows without the setup

No installs, no CUDA version roulette, no GPU sitting idle on your bill. Bring a workflow and run it in the browser.

Learn more