Civitai #143532220 Sunlight Boudoir: DesireX Krea 2 Chiaroscuro Portrait

Full Resolution

Generation Specs RTX 4070

Sampler euler

Scheduler simple

Steps 9

CFG Scale 1

Seed 741289

Target VRAM 12GB

VRAM 10GB - 12GB

GPU Hardware Compatibility

Optimal Ground Truth

Native GPU execution at full throughput with zero memory swapping.

DesireX (Krea 2 INT8 DiT) Civitai #143532220 Min 10GB VRAM 840×1120 RTX 4070 Verified

Sunlight Boudoir: DesireX Krea 2 Chiaroscuro Portrait

Pixel-exact reproduction of Civitai artwork #143532220 featuring DesireX Krea 2 DiT architecture, natural window blind chiaroscuro lighting, and Wan 2.1 16-channel VAE decoding.

Blueprint Summary RTX 4070 Verified

Reproducible ComfyUI workflow for Civitai #143532220 Sunlight Boudoir: DesireX Krea 2 Chiaroscuro Portrait using DesireX (Krea 2 INT8 DiT) at 840×1120 resolution. Requires minimum 10GB VRAM with sampler euler and scheduler simple (9 steps). Includes 1-click terminal model sync and canvas JSON graph.

Node Graph Pipeline 11 Total Nodes

Verify in Resolver

01 Load Models

UNETLoader

5 loaders (DiT, CLIP, VAE)

02 Adapters

LoraLoader

2 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

Model & Asset Setup 1-Click Script

Run in your ComfyUI root:

curl -fsSL https://decomfy.com/api/scripts/civitai-desirex-sunlight-boudoir-krea2.sh | bash

Positive Prompt

A young adult woman with fair skin and long blonde hair lying on soft bedding in a sunlit bedroom, photographed in an intimate close-up portrait orientation. She wears an elegant dark silk bikini set, a black choker necklace, and sheer black thigh-high stockings. Her dark red burgundy lips are slightly parted as she looks toward the camera with an expressive, captivating gaze. Strong natural morning sunlight streams in from the left window, casting dramatic linear shadow patterns from window blinds and golden highlights across her pale skin. In the background is a cozy modern bedroom with a minimalist gray dresser and warm beige carpet. High contrast chiaroscuro lighting, warm golden tones, natural skin texture, aesthetic smartphone selfie photography style, 8k raw photo.

Negative Prompt

blurry, low quality, distorted, extra limbs, ugly, deformed, oversaturated, painting, cartoon, doll, plastic

LoRA Adapter Stack 2 Adapters

RTX 4070 Calibrated Weights

01 Detailer-KREA2.safetensors +1

02 Krea2_TextFusion_Refusal_Reduction.safetensors +1

Required Models 5 Models

Disk Space Required: 17.5 GB (5 models · DiT/Base: 12.8 GB · Text Encoder: 4.4 GB · LoRAs: 39 MB · VAE: 254 MB)

models/diffusion_models/ 12.8 GB Civitai

BSSDesirexKrea2_x2INT8ConvrotFastest.safetensors DesireX Krea2 INT8 ConvRot fastest quantized DiT core

Get

models/text_encoders/ 4.4 GB HuggingFace

krea2_solordz_te_int8.safetensors Qwen3-VL 4B INT8 Text Encoder (Krea 2 Architecture)

Get

models/vae/ 254 MB HuggingFace

wan_2.1_vae.safetensors 16-channel VAE (prevents inverted contrast artifacts)

Get

models/loras/ 11.5 MB Civitai

Detailer-KREA2.safetensors Weight: 1.0 (Micro skin details & pore enhancer)

Get

models/loras/ 27.6 MB Civitai

Krea2_TextFusion_Refusal_Reduction.safetensors Weight: 1.0 (Refusal reduction & prompt compliance adapter)

Get

Field Notes RTX 4070 Benchmark

Benchmark: 12s on RTX 4070 (9.2 GB VRAM peak)

Ground-Truth Reproduction & White-Weapon Protocol: Reverse-engineered from Civitai #143532220 created with DesireX Krea2. Leverages the INT8 ConvRot fastest diffusion checkpoint combined with Qwen3-VL 4B text encoder and Wan 2.1 16-channel VAE (preventing inverted contrast artifacts). Features pre-wired Dual Prompt conditioning: Node A provides an instant compliant glamour boudoir lookbook, while Node B holds the standby prompt for 1-click local cable swapping. Benchmarked at 12s on RTX 4070.

Frequently Asked Questions FAQ

What GPU and VRAM are required to run Civitai #143532220 Sunlight Boudoir: DesireX Krea 2 Chiaroscuro Portrait?

This workflow requires a minimum of 10GB VRAM (recommended 12GB VRAM). Tested and verified on NVIDIA GeForce RTX 4070 (12GB VRAM) at 840x1120 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.

Which VAE and conditioning text encoders are required for Krea 2 DiT?

Krea 2 DiT single-stream models strictly require the 16-channel Wan2.1 VAE and Qwen3-VL text encoder. Using standard Flux or SDXL VAE causes inverted checkerboard artifacts. Also pair with RescaleCFG for high-contrast portrait clarity.