IMPORTANT — PLEASE READ BEFORE REPORTING CHARACTER LoRA ISSUES

The Krea2T Character LoRA — Image Only loader does NOT behave like a normal LoRA loader, and it is not intended to.

Comparing both loaders at 1.00 strength and assuming they should produce identical character likeness is an invalid comparison.

This node intentionally applies the character LoRA only to the image-token rows inside Krea2’s shared DiT blocks. It does not directly apply the character LoRA to the text-token rows.

Text and image still interact normally through Krea2’s native attention; the character LoRA itself is simply prevented from directly modifying the text side.

That behavior is intentional.

Strength Is Not 1:1 With a Normal LoRA Loader

Because the character LoRA is being applied through a more restricted mechanism, 1.00 on this loader is not equivalent to 1.00 on a conventional LoRA loader.

If your character likeness becomes weaker, increase the character LoRA strength above 1.00 until the desired likeness is recovered.

The node allows strengths above 1.00 for exactly this reason.

Why the Loader Works This Way

The refusal-reduction LoRA and the character LoRA have different jobs.

The refusal-reduction LoRA is intended to alter the model behavior responsible for refusal.

The character LoRA should provide the character appearance without unnecessarily becoming part of the prompt-interpretation path that the refusal-reduction LoRA is working against.

The entire point of this loader is therefore to apply the character LoRA as image influence rather than as a whole-entity influence across the full shared path.

Please Keep This in Mind Before Reporting an Issue

- Do not assume equal strength values between this loader and a normal LoRA loader represent equal influence

- Weaker likeness at 1.00 does not mean the loader is broken. Increase the character strength.

- This loader intentionally prevents direct character-LoRA contribution to text-token rows.

- This does not mean the character LoRA is “not working.”

- A difference in likeness does not prove that the likeness is located in the text encoder.

- If you specifically want the character LoRA applied conventionally across the normal model path, **use your normal LoRA loader instead.

Intended Separation

The design is simple:

Refusal-reduction LoRA → deals with refusal behavior

Character LoRA → supplies the visual character information

The goal is to keep the character LoRA from unnecessarily interfering with the interpretation path that the refusal-reduction LoRA is designed to modify.

Please understand what the node is designed to do before concluding that the behavior is a bug.

More importantly if you are using the Convrot version you MUST use this loader and enable Dynamic lora mode :

repo: https://github.com/BobJohnson24/ComfyUI-INT8-Fast

sample workflow : https://github.com/capitan01R/ComfyUI-Krea2T-Enhancer/blob/main/workflows/character_LoRA_wf.json

Krea2 Refusal-Reduction — v2 and v2 Full Rank

Okay, I am adding two versions this time: v2 and v2 Full Rank.

The goal is still the same: reduce refusal behavior while preserving as much of the model's existing visual knowledge and prompt adherence as possible. Getting a concept to appear initially is only part of it. I also want it to hold together through the denoise without drifting into something else or ending up with those weird textures.

There is also an important change to the setup if you use character LoRAs. Please read that part before loading everything together, because conflicts between them have been a recurring problem.

The Two Versions

v2

File: refusal_reduction_v2.safetensors

This is the smaller rank-64 version, approximately 30 MB. It includes updates to the four TextFusion blocks, the tap projector, and the two external text MLP layers that prepare the conditioning for the image transformer.

v2 Full Rank

File: refusal_reduction_v2_full_rank.safetensors

This is the larger full-rank export, approximately 2.6 GB. It covers the same main conditioning components and also includes selected normalization adjustments.

The matrix updates are stored at their full available rank in FP32, which accounts for the much larger file. “Full Rank” describes how those updates are represented; the changes are still confined to the text-conditioning components covered by this release.

Load one version only. Start at strength 1.00.

What These Versions Change

Krea2 takes several representations of your prompt from its text encoder and processes them before they reach the image transformer. These releases modify that processing, with the aim of making the model less likely to suppress the requested concept.

Both versions include:

• The two TextFusion layerwise blocks.

• The two TextFusion refiner blocks.

• The projector that combines the twelve text-encoder representations.

• The two external text MLP layers that feed the resulting conditioning into the image transformer.

The earlier description saying the projector and external text MLPs were excluded does not apply to these two files.

Neither file contains updates to the image-transformer blocks. They change the conditioning those blocks receive, which can still change the resulting image substantially.

The purpose remains refusal reduction. I am not aiming to impose a character, style, or preferred aesthetic with these releases.

Using a Character LoRA? This Loader Is Required

If you combine either version with a character LoRA, use Krea2T Character LoRA — Image Only to load the character LoRA.

One thing I have repeatedly run into is a character LoRA pulling an otherwise working setup away from the requested concept. You get the likeness, but other parts of the prompt start behaving differently or become less reliable.

Simply excluding weights from the dedicated text-processing layers does not fully separate their influence. Krea2 also has shared layers that process both the prompt and the image. A character LoRA applied normally to those layers affects both.

That is why I added this loader. It filters the supported character weights and applies their direct updates only to image tokens. The refusal-reduction LoRA stays active through its own loader.

Text and image still interact through the model's normal attention, so some interaction remains. The loader targets that direct overlap rather than treating every conflict as a reason to keep increasing LoRA strength.

Connect It Like This

Krea2 model
↓
Regular LoRA loader — refusal_reduction_v2 OR refusal_reduction_v2_full_rank
↓
Krea2T Character LoRA — Image Only — your character LoRA
↓
Sampler

• Load the refusal-reduction LoRA through your regular model LoRA loader.

• Load the character LoRA through the Image Only loader. Replace its previous loader; do not load the same character twice.

• Do not put the refusal-reduction LoRA into the Image Only loader. Its filtering excludes the conditioning weights these releases use.

• If you are not using a character LoRA, connect the refusal-reduction loader directly to the rest of your workflow.

Still a Work in Progress

I am continuing to check how consistently this holds up across prompts, seeds, and different LoRA combinations. The objective is useful refusal reduction with the concept intact through generation. I am not calling every changed image a success.

Buy me a coffee :)

Responsibility and Output Disclaimer

This LoRA modifies the model's text-conditioning behavior but does not determine, authorize, endorse, supervise, or control the images produced by individual users, prompts, workflows, checkpoints, samplers, or third-party software.

Generated outputs remain dependent on the base model, user-provided conditioning, inference configuration, and the surrounding generation pipeline. The author assumes no responsibility or liability for what users choose to generate, how generated material is used, or whether any output complies with applicable laws, platform rules, licensing terms, or third-party rights.

Users are solely responsible for their prompts, generated outputs, distribution decisions, and use of this LoRA.