Civitai #145918769: Touko Sakurai (My Senpai is Annoying) - Unholy Desire Mix Illustrious
Full Resolution
Generation Specs RTX 4070
Sampler dpmpp_2m
Scheduler karras
Steps 28
CFG Scale 5
Seed 3753469294
Target VRAM 12GB
VRAM 8GB - 12GB
GPU Compatibility
Optimal Performance

Runs entirely in VRAM with no memory swapping.

Unholy Desire Mix - Sinister Aesthetic (Illustrious) Civitai #145918769 Min 8GB VRAM 1072×1920 RTX 4070 Verified

Touko Sakurai (My Senpai is Annoying) - Unholy Desire Mix Illustrious

High-fidelity anime screencap and lookbook portrait of Touko Sakurai deconstructed from Civitai #145918769 using Unholy Desire Mix v7.0 and a 4-LoRA stack (Character, Detailer IL, 748cm Style, Niji Semi Realism).

Blueprint Summary RTX 4070 Verified

Reproducible ComfyUI workflow for Civitai #145918769: Touko Sakurai (My Senpai is Annoying) - Unholy Desire Mix Illustrious using Unholy Desire Mix - Sinister Aesthetic (Illustrious) at 1072×1920 resolution. Requires minimum 8GB VRAM with sampler dpmpp_2m and scheduler karras (28 steps). Includes 1-click terminal model sync and canvas JSON graph.

Pipeline Flow 14 Nodes
Open in Resolver
01 Load Models
CheckpointLoaderSimple
5 loaders (DiT, CLIP, VAE)
02 Adapters
LoraLoader
4 adapters
03 Conditioning
CLIPTextEncode
2 prompt encodings
04 KSampler
KSampler
DiT latent denoising
05 Decode & Save
VAEDecode
Latent to pixel space

Execution DAG Topology Interactive Visualizer

Drag to pan · Scroll to zoom · Hover wires
Node Graph 0 nodes
MODEL CLIP LATENT VAE IMAGE

Model & Asset Setup Setup Script

Run in your ComfyUI root:

curl -fsSL https://decomfy.com/api/scripts/civitai-touko-sakurai-unholy-desire-illustrious.sh | bash

Positive Prompt

SemiFrealism, Jeddtl02, 748cmstyle, (masterpiece, best quality, very aesthetic, amazing quality, absurdres, newest), 1girl, solo, touko sakurai, black hair, sidelocks, medium hair, mole under eye, swept bangs, mature female, purple eyes, round eyewear glasses, cheerful smile, looking at viewer, hand in pocket, hand up waving, wearing stylish blue denim overalls over a cream ribbed knit tank top, clean collarbone, bare shoulders, cowboy shot, simple pastel yellow background, crisp black ink outlines, stepped cel shading, vibrant colors, anime screencap

Negative Prompt

unclothed, disrobed, revealing clothes, lowres, blurry, bad anatomy, deformed hands, extra fingers, low quality, artifacts, distorted proportions, text, watermark, signature, score_4, score_5, score_6
Prompt Customization: This workflow is pre-wired and calibrated. Swap this prompt with your own character, subject, or style without breaking the pipeline.

LoRA Adapter Stack 4 Adapters

RTX 4070 Calibrated Weights
01 touko-sakurai-s1-illustriousxl-lora-nochekaiser.safetensors +1
02 DetailerILv2-000008.safetensors +0.6
03 748cmSDXL.safetensors +0.8
04 Niji_Semi_realism_F_N_R_epoch_10.safetensors +0.8

Required Models 5 Models

Required Storage: 7.5 GB (5 models · DiT/Base: 6.6 GB · LoRAs: 897 MB)
models/checkpoints/ 6.62 GB Civitai
unholyDesireMixSinister_v70.safetensors Official Unholy Desire Mix - Sinister Aesthetic v7.0 (Illustrious XL Checkpoint)
models/loras/ 218 MB Civitai
touko-sakurai-s1-illustriousxl-lora-nochekaiser.safetensors Touko Sakurai (My Senpai is Annoying) Anime S1 Character LoRA by nochekaiser
models/loras/ 218 MB Civitai
DetailerILv2-000008.safetensors Detailer IL v2 by Jedit (Trigger: Jeddtl02 - Micro-details, stitching, eye reflections)
models/loras/ 243 MB Civitai
748cmSDXL.safetensors 748cm Style for Illustrious (Trigger: 748cmstyle - Illustrative brushwork and shading)
models/loras/ 218 MB Civitai
Niji_Semi_realism_F_N_R_epoch_10.safetensors Niji Semi Realism v4.0 for Illustrious (Trigger: SemiFrealism - Dimensional painterly rendering)

Field Notes RTX 4070 Benchmark

Benchmark: Rendered on NVIDIA GeForce RTX 4070 (12GB VRAM, CUDA 12.4): 28+26 steps DPM++ 2M Karras @ CFG 5.0/6.0, 1072x1920 (Hi-Res 2x), 8.9 GB peak VRAM, 69s execution time.
1. CIVITAI PROVENANCE & HIDDEN LORA DECONSTRUCTION: Forensic analysis of Civitai image #145918769 by creator AlexAlexHen revealed an intricate multi-LoRA recipe leveraging Unholy Desire Mix - Sinister Aesthetic v7.0 (Model #1307857, Version #2403075). Beyond the character LoRA (Touko Sakurai Anime S1, Version #1753519), the creator activated three distinct style LoRAs via trigger tokens in Forge WebUI: Detailer IL v2 (Jeddtl02, Version #1736373), 748cm Style (748cmstyle, Version #1056404), and Niji Semi Realism v4.0 (SemiFrealism, Version #2854725). Together, this 4-LoRA stack elevates flat anime rendering into rich semi-realistic painterly dimensions with micro-stitching detail on garments and nuanced iris highlights. 2. WHITE-WEAPON SFW RE-ENGINEERING & UNIVERSAL CLEAN NEGATIVES: The raw Civitai artwork relied on unconstrained anatomical sliders (Civitai NSFW Level 8). Under the White-Weapon Protocol ('Tẩy Trắng'), positive conditioning anchors Touko Sakurai in stylish blue denim overalls layered over a cream ribbed knit tank top, preserving the signature cowboy framing, waving gesture, and cheerful pastel yellow backdrop while eliminating all adult nudity. Negative conditioning strictly enforces Universal Clean Negatives combined with Illustrious quality tags (score_4, score_5, score_6) to ensure 100% Google SafeSearch indexation compliance. 3. HARDWARE BENCHMARKS & 4-LORA SIGNAL FLOW ON RTX 4070: Executed natively on external GPU host wingpu (NVIDIA GeForce RTX 4070 12GB VRAM, CUDA 12.4). The pipeline feeds Unholy Desire Mix v7.0 through CLIPSetLastLayer (-2) and chains all 4 LoRAs sequentially in memory. Initial txt2img generates at 540x960 (28 steps DPM++ 2M Karras @ CFG 5.0), followed by 2.0x bicubic LatentUpscaleBy and a Hi-Res refinement pass (26 steps @ CFG 6.0, denoise 0.55), delivering a razor-sharp 1072x1920 output. Peak VRAM was measured at 8.9 GB with a total execution time of 69 seconds.

Frequently Asked Questions FAQ

What GPU and VRAM are required to run Civitai #145918769: Touko Sakurai (My Senpai is Annoying) - Unholy Desire Mix Illustrious?

This workflow requires a minimum of 8GB VRAM (recommended 12GB VRAM). Tested and verified on NVIDIA GeForce RTX 4070 (12GB VRAM) at 1072x1920 resolution.

How do I resolve missing custom nodes for this workflow?

You can drop the workflow JSON into our client-side Missing Node Auto-Resolver at https://decomfy.com/resolve/ to detect missing nodes and generate install commands, or run the 1-click terminal setup script provided below.

What text encoders and VAE does Qwen-Image 2.1 Uncensored require?

Qwen-Image 2.1 Uncensored uses GGUF Q4_K_M quantized DiT weights paired with the Heretic uncensored text encoder, bypassing safety refusals while maintaining precise prompt adherence.

Related ComfyUI Blueprints

View all 56 workflows
Krea 2 DiT
Fast Turbo / DMD2 #144401284

Dahlia Editorial Studio Portrait - Krea 2 Turbo

8st · euler 832×1248 12GB
Flux.1 Dev 16fps Video
Photorealism Portrait

The Godfather Part II: English Bulldog Edition - Enemies Closer 2K Video

25st · euler 1440×2560 12GB
Flux.1 Dev 16fps Video
Photorealism Portrait

The Godfather: English Bulldog Edition - 2K MiniMax H3 Cinematic Video

25st · euler 1440×2560 12GB
Action completed