Civitai #144982792: Candid Flash Loft Portrait - SNOFS & Krea 2 Turbo

Full Resolution

Generation Specs RTX 4070

Sampler euler

Scheduler simple

Steps 10

CFG Scale 1

Seed 468292559530759

Target VRAM 12GB

VRAM 8GB - 12GB

GPU Compatibility

Optimal Performance

Runs entirely in VRAM with no memory swapping.

Krea 2 Turbo (Official Comfy-Org) Civitai #144982792 Min 8GB VRAM 832×1248 RTX 4070 Verified

Candid Flash Loft Portrait - SNOFS & Krea 2 Turbo

Authentic flash-lit night snapshot portrait featuring candid facial expressions, micro skin texture, and vintage loft interior aesthetics synthesized with SNOFS v1.5 on Krea 2 Turbo.

Blueprint Summary RTX 4070 Verified

Reproducible ComfyUI workflow for Civitai #144982792: Candid Flash Loft Portrait - SNOFS & Krea 2 Turbo using Krea 2 Turbo (Official Comfy-Org) at 832×1248 resolution. Requires minimum 8GB VRAM with sampler euler and scheduler simple (10 steps). Includes 1-click terminal model sync and canvas JSON graph.

Pipeline Flow 10 Nodes

Open in Resolver

01 Load Models

UNETLoader

4 loaders (DiT, CLIP, VAE)

02 Adapters

LoraLoader

1 adapter

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 Setup Script

Run in your ComfyUI root:

curl -fsSL https://decomfy.com/api/scripts/civitai-flash-loft-portrait-krea2.sh | bash

Positive Prompt

A high-contrast flash-lit night snapshot of a chic 22-year-old woman smiling warmly towards the camera, wearing an oversized black tailored wool blazer over a crisp white cotton tee, sitting on a vintage leather sofa in a dimly lit loft apartment at night. Direct harsh on-camera flash, sharp specular highlights, authentic candid expression, natural skin texture with subtle pores, deep atmospheric flash shadows in the background, 35mm film snapshot realism.

Negative Prompt

unclothed, disrobed, revealing clothes, lowres, blurry, bad anatomy, deformed hands, extra fingers, low quality, artifacts, distorted proportions, text, watermark, cartoon, 3d render, plastic skin

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 1 Adapters

RTX 4070 Calibrated Weights

01 SNOFS (Sex, Nudes, Other Fun Stuff) v1.5 +1

Required Models 4 Models

Required Storage: 19.2 GB (4 models · DiT/Base: 12.6 GB · Text Encoder: 4.9 GB · LoRAs: 1.5 GB · VAE: 253 MB)

models/diffusion_models/ 12.56 GB HuggingFace

krea2_turbo_int8_convrot.safetensors Official Krea 2 Turbo INT8 ConvRot Diffusion Transformer core

Get

models/text_encoders/ 4.88 GB HuggingFace

qwen3vl_4b_fp8_scaled.safetensors Official Qwen3-VL 4B FP8 Scaled Text Encoder (CLIPLoader type: krea2)

Get

models/vae/ 253 MB HuggingFace

qwen_image_vae.safetensors Official 16-channel Video & Image VAE

Get

models/loras/ 1.49 GB HuggingFace (Mirror)

snofs_krea_v1_5.safetensors Weight: 1.0 (SNOFS v1.5 candid flash photography snapshot LoRA for realistic skin pores, direct flash lighting, and unposed facial expressions)

Get

Field Notes RTX 4070 Benchmark

Benchmark: Rendered on NVIDIA GeForce RTX 4070 (12GB VRAM, CUDA 12.4)

1. ARCHITECTURAL REASONING & COMPOSITION GEOMETRY: Deconstructs the raw, amateur flash photography snapshot aesthetic pioneered by the SNOFS (Sex, Nudes, Other Fun Stuff) v1.5 LoRA on Krea 2 Turbo. While community usage frequently channels SNOFS into explicit adult photography, this reproduction leverages White-Weapon semantic sanitization to extract its true technical superpower: authentic micro skin texture, specular skin highlights, unposed facial geometry, and high-contrast on-camera flash shadow falloff. Anchoring the positive prompt with structured garments (oversized tailored wool blazer over a crisp crewneck tee) stabilizes latent spatial attention and prevents anatomical degradation, achieving gallery-grade candid portrait realism. 2. CABLE ROUTING & SIGNAL TOPOLOGY: Single-pass sequential LoRA injection topology routing from Krea 2 Turbo INT8 UNET (krea2_turbo_int8_convrot.safetensors) and dedicated Qwen3-VL 4B text encoder (qwen3vl_4b_fp8_scaled.safetensors). The SNOFS v1.5 adapter modifies both model weights and text encoder embeddings at full strength (1.0). Conditioning passes into a single deterministic positive conditioning node wired directly to KSampler, adhering strictly to the WYSIWYG single-node invariant with zero standby or muted nodes. Latents are decoded through the native 16-channel qwen_image_vae.safetensors. 3. HARDWARE BENCHMARKS & CONSUMER GPU FOOTPRINT: Synthesized and verified on host NVIDIA GeForce RTX 4070 (12GB VRAM, CUDA 12.4). Peak VRAM consumption clocked at 3.2 GB during the 10-step Euler generation at 832x1248 resolution, consuming a mere 26% of available VRAM. Total generation latency completed in 81.2 seconds, confirming frictionless deployment for 8GB and 12GB consumer GPUs without CPU paging.

Frequently Asked Questions FAQ

What GPU and VRAM are required to run Civitai #144982792: Candid Flash Loft Portrait - SNOFS & Krea 2 Turbo?

This workflow requires a minimum of 8GB VRAM (recommended 12GB VRAM). Tested and verified on NVIDIA GeForce RTX 4070 (12GB VRAM) at 832x1248 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.