Civitai #144666145: Bedroom Candid Snapshot - Realistic Flash Photography & Krea 2 Turbo

Full Resolution

Generation Specs RTX 4070

Sampler euler

Scheduler simple

Steps 10

CFG Scale 1

Seed 2626861826

Target VRAM 12GB

VRAM 8GB - 12GB

GPU Hardware Compatibility

Optimal Ground Truth

Native GPU execution at full throughput with zero memory swapping.

Krea 2 Turbo (Official Comfy-Org) Civitai #144666145 Min 8GB VRAM 832×1248 RTX 4070 Verified

Bedroom Candid Snapshot - Realistic Flash Photography & Krea 2 Turbo

Hyper-realistic amateur bedroom snapshot rendered on Krea 2 Turbo. Leverages the Realistic Snapshot LoRA with direct flash photography, unposed candid posture, and authentic micro-textured skin detail.

Blueprint Summary RTX 4070 Verified

Reproducible ComfyUI workflow for Civitai #144666145: Bedroom Candid Snapshot - Realistic Flash Photography & 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.

Node Graph Pipeline 12 Total Nodes

Verify in Resolver

01 Load Models

UNETLoader

5 loaders (DiT, CLIP, VAE)

02 Adapters

LoraLoader

2 adapters

03 Conditioning

CLIPTextEncode

3 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-bedroom-candid-snapshot-krea2.sh | bash

Positive Prompt

Candid amateur smartphone photo, direct harsh flash photography, cozy modern bedroom at night. A young attractive woman with long brunette hair wearing soft white ribbed tank top and casual lounge shorts, relaxing leisurely on a king-sized bed with crisp white cotton sheets and a dark grey tufted headboard. Natural smiling expression, genuine candid moment, authentic unposed posture, textured realistic skin, slight film grain, high dynamic range flash lighting, natural bedroom interior with warm ambient background light, door and light switch visible in background, 4k sharp details, photorealistic editorial snapshot.

Negative Prompt

deformed, bad anatomy, deformed hands, extra fingers, missing fingers, lowres, blurry, distorted proportions, low quality, artifacts, watermark, signature, cartoon, 3d render, plastic skin, CGI, airbrushed, overexposed blown highlights

Custom Prompt Freedom: This workflow is 100% pre-wired. The canvas includes a secondary standby conditioning node (muted mode: 2). When importing into ComfyUI, you can unmute it or freely write any custom prompt, characters, or aesthetic styles. The underlying model checkpoint, LoRA adapter stack, and sampling pipeline will faithfully execute your vision.

LoRA Adapter Stack 2 Adapters

RTX 4070 Calibrated Weights

01 Realistic Snapshot (Z-Image-Turbo + Krea 2) +1

02 Krea2 Refusal Reduction V2 (Full Rank) +1

Required Models 5 Models

Disk Space Required: 21.7 GB (5 models · DiT/Base: 12.6 GB · Text Encoder: 4.9 GB · LoRAs: 4.0 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 Civitai

RealisticSnapshotKrea2V2.safetensors Weight: 1.0 (Realistic Snapshot adapter for authentic amateur flash camera aesthetics)

Get

models/loras/ 2.48 GB HuggingFace (Mirror)

refusal_reduction_v2_full_rank.safetensors Weight: 1.0 (Full-rank Refusal Reduction LoRA for photorealistic skin texture alignment)

Get

Field Notes RTX 4070 Benchmark

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

1. ARCHITECTURAL REASONING & REALISTIC SNAPSHOT INTEGRATION: Krea 2 Turbo natively achieves remarkable anatomical fidelity, but default generations often skew toward overly polished commercial glamour. Injecting the Realistic Snapshot LoRA (weight 1.0) shifts the visual distribution toward authentic amateur smartphone photography, introducing high-contrast direct flash, realistic skin pores, and natural indoor ambient falloff. 2. TEXT ENCODER & REFUSAL MITIGATION: Driven by Qwen3-VL 4B FP8 Scaled text encoder via native CLIPLoader (type: krea2). Chaining the full-rank Refusal Reduction LoRA upstream ensures complex candid descriptions and natural skin cues compile without semantic filtering or degradation. 3. HARDWARE BENCHMARKS & TURBO EFFICIENCY: Rendered on NVIDIA GeForce RTX 4070 (12GB VRAM). 10 inference steps of Krea 2 INT8 ConvRot complete in ~46 seconds at 832x1248, maintaining steady memory headroom under 10.5GB.

Frequently Asked Questions FAQ

What GPU and VRAM are required to run Civitai #144666145: Bedroom Candid Snapshot - Realistic Flash Photography & 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.