---

A ComfyUI node that automatically analyzes your LoRA stack and selects the best merge strategy per weight group — diff-based merging, TIES conflict resolution, per-prefix adaptive decisions, SVD patch compression, and auto-tuned parameters. Two nodes: LoRA Stack (build input) and LoRA Optimizer (analyze + merge).

The Problem

Stacking LoRAs in ComfyUI adds their effects together. On distilled/turbo models (SDXL-Turbo, LCM, Lightning, Flux-schnell), the accumulated effect exceeds what the model can handle, causing overexposure, color blowout, and artifacts.

model += lora1_effect x strength1
model += lora2_effect x strength2
total effect = strength1 + strength2  -->  easily exceeds 1.0

The optimizer solves this by computing full weight diffs, detecting sign conflicts per weight group, and merging each group with its optimal strategy.

Nodes

LoRA Stack

Builds a list of LoRAs for the optimizer. Chain multiple Stack nodes to add any number of LoRAs.

Inputs: LoRA selector, strength, optional previous LORA_STACK

Outputs: LORA_STACK

---

LoRA Optimizer

The auto-optimizer. Takes a LORA_STACK, analyzes the LoRAs, and automatically selects the best merge mode and parameters per weight group. Outputs the merged result plus a detailed analysis report with a block strategy map.

Also accepts standard tuple-format stacks (lora_name, model_strength, clip_strength) from Efficiency Nodes, Comfyroll, and similar packs.

Uses a two-pass streaming architecture for low memory usage:

• Pass 1 (Analysis): Computes weight diffs per prefix, samples conflict and magnitude statistics per prefix, then discards the diffs. Only lightweight scalars are kept.
• Pass 2 (Merge): Recomputes diffs per prefix, looks up that prefix's conflict data, picks the optimal strategy for it, and merges. Each prefix is freed after merging. Non-TIES patches are SVD-compressed to low-rank by default.

Peak memory is ~one prefix at a time (~260MB) regardless of LoRA count or model size. GPU-accelerated on both passes.

Per-Prefix Adaptive Merge

The key insight: two LoRAs may overlap in some model blocks but not others. A face LoRA and a style LoRA might only conflict in attention layers 4-7, while the rest of the model is touched by only one of them.

Instead of picking one global strategy (which either wastes TIES trimming on non-overlapping blocks or misses real conflicts), the optimizer decides per weight prefix:

Condition	Strategy
Only 1 LoRA touches this prefix	weighted_sum — full strength, no dilution
2+ LoRAs, sign conflict <= 25%	weighted_average — compatible, simple merge
2+ LoRAs, sign conflict > 25%	ties — resolve conflicts with trim/elect/merge
Magnitude ratio > 2x at prefix	total sign method (stronger LoRA dominates)
Magnitude ratio <= 2x at prefix	frequency sign method (equal votes)

This means non-overlapping regions keep 100% of their LoRA's effect, while genuinely conflicting regions get proper TIES resolution.

Optimization Modes

Mode	Behavior
per_prefix (default)	Each weight group picks its own strategy based on local conflict data
global	Single strategy for all prefixes (original behavior)
weighted_sum_only	Forces simple weighted sum everywhere — no TIES, no averaging. Combined with patch compression, all patches are fully compressible with zero quality loss

Block Strategy Map

The analysis report includes a visual block-by-block map showing what strategy was used and why:

--- Block Strategy Map ---
  input_blocks.0   ====  sum  1 LoRA (6x)
  input_blocks.4   ----  avg  12% conflict (6x)
  middle_block.1   ####  TIES 42% conflict (6x)
  output_blocks.3  ----  avg  8% conflict (6x)
  output_blocks.8  ====  sum  1 LoRA (6x)
  Legend: ==== sum (single LoRA)  ---- avg (compatible)  #### TIES (conflict)

What It Analyzes

• Per-LoRA metrics (rank, key count, effective L2 norms)
• Pairwise sign conflict ratios per prefix (sampled for efficiency)
• Magnitude distribution per prefix
• Key overlap between LoRAs

TIES Merging

The optimizer automatically selects TIES-Merging (Trim, Elect Sign, Disjoint Merge — Yadav et al., NeurIPS 2023) on prefixes where sign conflicts are detected between LoRAs.

SVD Patch Compression

After merging, full-rank diff patches consume ~128x more RAM than standard LoRA patches (64MB vs 0.5MB per key for a 4096x4096 weight). The optimizer re-compresses merged patches to low-rank via truncated SVD, dramatically reducing post-merge RAM.

Mode	What gets compressed	Quality	RAM savings
non_ties (default)	weighted_sum and weighted_average prefixes only	Lossless — sum of input ranks preserves all merge information	~32x on compressed prefixes
all	Everything including TIES	Lossy on TIES prefixes — nonlinear ops (trim, sign election) produce full-rank results that can't be perfectly captured	~32x on all prefixes
disabled	Nothing	No loss	No savings

The compression rank is automatically computed as the sum of all input LoRA ranks. For example, 3 rank-32 LoRAs produce a rank-96 compressed patch — enough to represent the full merge without quality loss on linear operations.

Tip: For video models (LTX2, Wan, etc.) with high RAM usage, use weighted_sum_only + non_ties (or all). Every patch gets losslessly compressed with minimal RAM footprint.

Auto-Strength

When auto_strength is set to enabled, the optimizer automatically reduces per-LoRA strengths before merging to prevent overexposure from stacking. This is especially useful on distilled/turbo models where 2+ LoRAs at full strength cause blown-out results even with optimal merge mode selection.

The algorithm uses L2-aware energy normalization: it measures each LoRA's actual weight magnitude (L2 norms collected during analysis) and scales all strengths so the total combined energy matches what the strongest single LoRA would contribute alone.

Scenario	Result
2 equal LoRAs at strength 1.0	Each reduced to ~0.71 (1/sqrt(2))
1 strong + 1 weak LoRA	Proportional reduction, preserving ratio
Single LoRA	No change (scale = 1.0)
auto_strength disabled	No adjustment (default)

Your original strength ratios are always preserved — the algorithm only scales them down uniformly.

Memory Options

Option	Default	Effect
cache_patches	enabled	Cache merged patches in RAM for faster re-execution. Disable to free RAM after merge (recommended for video models)
compress_patches	non_ties	SVD re-compression of merged patches (see above)
free_vram_between_passes	disabled	Release GPU cache between analysis and merge passes. Lowers peak VRAM at negligible speed cost

Inputs / Outputs

Inputs: MODEL, CLIP (optional), LORA_STACK, output strength, clip strength multiplier, auto strength, optimization mode, cache patches, compress patches, free VRAM between passes.

Outputs: MODEL, CLIP, STRING (analysis report)

Example Report

==================================================
LORA OPTIMIZER - ANALYSIS REPORT
==================================================

--- Per-LoRA Analysis ---
  style_lora.safetensors:
    Strength: 1.0
    Keys: 192
    Avg rank: 64
    L2 norm (mean): 0.0847
  detail_lora.safetensors:
    Strength: 0.8
    Keys: 192
    Avg rank: 32
    L2 norm (mean): 0.0423

--- Auto-Strength Adjustment ---
  style_lora.safetensors: 1.0 -> 0.7071
  detail_lora.safetensors: 0.8 -> 0.5657
  Scale factor: 0.7071
  Method: L2-aware energy normalization

--- Pairwise Analysis ---
  style_lora.safetensors vs detail_lora.safetensors:
    Overlapping positions: 89420
    Sign conflicts: 31297 (35.0%)

--- Collection Statistics ---
  Total LoRAs: 2
  Total unique keys: 196
  Avg sign conflict ratio: 35.0%
  Magnitude ratio (max/min L2): 2.00x

--- Auto-Selected Parameters ---
  Merge mode: ties
  Density: 0.42
  Sign method: frequency
  (global fallback — each prefix uses its own parameters)

--- Per-Prefix Strategy ---
  weighted_sum (single LoRA):        28 prefixes (14%)
  weighted_average (low conflict):  120 prefixes (61%)
  ties (high conflict):              48 prefixes (24%)
  Total:                            196 prefixes

--- Block Strategy Map ---
  input_blocks.0   ====  sum  1 LoRA (6x)
  input_blocks.1   ====  sum  1 LoRA (6x)
  input_blocks.4   ----  avg  12% conflict (6x)
  input_blocks.5   ####  TIES 38% conflict (6x)
  middle_block.1   ####  TIES 42% conflict (6x)
  output_blocks.3  ----  avg  15% conflict (6x)
  output_blocks.8  ====  sum  1 LoRA (6x)
  Legend: ==== sum (single LoRA)  ---- avg (compatible)  #### TIES (conflict)

--- Reasoning ---
  Sign conflict ratio 35.0% > 25% threshold -> TIES mode selected
    TIES resolves sign conflicts via trim + elect sign + disjoint merge
  Auto-density estimated at 0.42 from magnitude distribution
  Magnitude ratio 2.00x <= 2x -> 'frequency' sign method (equal voting)
    Similar-strength LoRAs get equal votes

--- Merge Summary ---
  Keys processed: 196
  Model patches: 168
  CLIP patches: 28
  Output strength: 1.0
  CLIP strength: 1.0

==================================================

Connect the STRING output to a Show Text node to see the report in ComfyUI.

Structural LoRAs: Do not put distillation LoRAs (LCM, Lightning, Turbo, Hyper) or DPO LoRAs in the optimizer stack. These LoRAs modify the model's fundamental behavior (denoising process, preference alignment) and their weights are precisely calibrated — merging them with style LoRAs can break their training. Apply them via a standard Load LoRA node upstream, then feed only your style/character LoRAs into the optimizer.

Limitation: The optimizer only analyzes LoRAs in its own stack. It cannot see LoRA patches applied by upstream nodes (Load LoRA, etc.) — those stack additively on top of the optimizer's output. Fully baked merges (safetensors checkpoints) are indistinguishable from base weights and cannot be detected.

Installation

ComfyUI Manager

Search for "LoRA Optimizer" in ComfyUI Manager and install.

Manual

cd ComfyUI/custom_nodes/
git clone https://github.com/ethanfel/ComfyUI-LoRA-Optimizer.git

Restart ComfyUI. Both nodes appear under the loaders/lora category.

Compatibility

• Models: SD 1.5, SDXL, Flux, and other architectures supported by ComfyUI
• LoRA formats: Standard LoRA, LoCon, diffusers formats
• Flux sliced weights: Handled correctly (linear1_qkv offsets)
• Stack formats: Native LoRA Stack dicts, plus standard tuples from Efficiency Nodes / Comfyroll

Credits

• Originally based on ComfyUI-ZImage-LoRA-Merger by DanrisiUA
• Per-prefix adaptive approach inspired by comfyUI-Realtime-Lora by shootthesound (per-block LoRA analysis)
• Thanks to Scruffy and Ramonguthrie for suggesting the per-block analysis approach
• TIES-Merging: Yadav et al., NeurIPS 2023

License

MIT License - see LICENSE.