isiwork GitHub avatar

ComfyUI-Krea-LoRA-Tools

isiwork

A ComfyUI custom node pack for applying per-block weight adjustments to LoRAs on Krea 2's DiT model and for merging multiple LoRAs into a single .safetensors file.

Stars

12

7-day growth

No data

Forks

1

Open issues

0

License

MIT

Last updated

2026-07-12

AI repository intelligence
FR-AI / ANALYSIS

Why it is worth attention

It fills a gap by providing fine-grained control over LoRA activation per DiT block (0–27) and a simple pipeline to merge LoRAs with concat or SVD, a feature missing from many basic LoRA tools.

Who it is for

  • ComfyUI users working with Krea 2 models
  • LoRA creators who need per-block strength tuning
  • Users wanting to combine multiple LoRAs into one file
  • AI artists experimenting with diffusion model fine-tuning

Use cases

  • Adjust the influence of a LoRA across different DiT blocks during image generation
  • Merge several LoRAs with weighted strength into a single compact file
  • Convert a single LoRA to a different rank using SVD without changing content
  • Preserve a merge recipe as metadata for reproducibility

Strengths

  • Block-level weight specification using range or explicit list syntax
  • Two merge modes: lossless concat and rank-reducing SVD with configurable rank
  • Records merge source, strengths, and mode as metadata in the output file
  • Supports both ai-toolkit/diffusers and kohya LoRA naming conventions

Considerations

  • Only designed for Krea 2's DiT architecture; not tested on other models
  • Requires ComfyUI as the host environment
  • No built-in preview or validation of block weight effects before generation

README quick start

使用するノード

Related repositories

Similar projects matched by category, topics, and programming language.

lopopolo
Featured
lopopolo GitHub avatar

harness-engineering

Harness Engineering is a methodology for improving coding agent outputs by carefully crafting the environment around them—providing curated context, tools, and executable constraints that encode an organization’s nonfunctional requirements and cumulative lessons.

AI & Machine LearningAI Agents
2,390
slvDev
Featured
slvDev GitHub avatar

esp32-ai

A 28.9 million parameter language model runs on an $8 ESP32-S3 microcontroller entirely on-device, generating simple stories at about 9.5 tokens per second.

AI & Machine LearningLarge Language Models
1,960
littledivy
Featured
littledivy GitHub avatar

mimic

mimic captures traffic from any iOS or web app and automatically generates a Python client library that lets you call the app's API like a regular library.

AI & Machine Learning
1,482