Civitai #143442077 Reproduction: Shiba Inu Izakaya Yakitori Chef

Full Resolution

Generation Specs RTX 4070

Sampler euler

Scheduler simple

Steps 25

CFG Scale 1

Seed 91823

Target VRAM 12GB

VRAM 10GB - 12GB

GPU Hardware Compatibility

Optimal Ground Truth

Native GPU execution at full throughput with zero memory swapping.

Flux.1 Dev GGUF (Q4_K_S) Civitai #143442077 Min 10GB VRAM 768×1024 RTX 4070 Verified

Reproduction: Shiba Inu Izakaya Yakitori Chef

Authentic izakaya culinary artwork reproduction of viral Civitai #143442077. Benchmarked on RTX 4070 using Flux.1 Dev GGUF (Q4_K_S) with DualCLIP (ViT-L + T5-XXL) and binchotan charcoal smoke dynamics.

Blueprint Summary RTX 4070 Verified

Reproducible ComfyUI workflow for Civitai #143442077 Reproduction: Shiba Inu Izakaya Yakitori Chef using Flux.1 Dev GGUF (Q4_K_S) at 768×1024 resolution. Requires minimum 10GB VRAM with sampler euler and scheduler simple (25 steps). Includes 1-click terminal model sync and canvas JSON graph.

Node Graph Pipeline 10 Total Nodes

Verify in Resolver

01 Load Models

UnetLoaderGGUF

3 loaders (DiT, CLIP, VAE)

02 Conditioning

CLIPTextEncode

3 prompt encodings

03 KSampler

KSampler

DiT latent denoising

04 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-shiba-yakitori-chef-flux-dev.sh | bash

Positive Prompt

An authentic raw candid photograph of a real Shiba Inu dog working as a yakitori chef in a cramped dark Tokyo izakaya alley at night. The Shiba Inu is standing behind a smoking iron yakitori grill with glowing binchotan charcoal beneath. It has natural textured canine fur with visible individual hairs, slightly greasy and smoky from the grill, and focused squinting dark eyes looking down at the cooking skewers. It wears only a weathered white cotton hachimaki headband with a faded red sun circle and kanji, no other clothing. Its real furry paws hold wooden skewers of chicken pieces (negima) with charred edges, bubbling soy tare glaze, and green onions. Heavy white smoke plumes billow from the hot coals, catching the warm red lantern light. Authentic Japanese izakaya, wooden slat walls, red lantern with calligraphy, shallow depth of field, 35mm photography, natural film grain, raw photo, f/1.8 lens.

Negative Prompt

cartoon, 3d render, illustration, plush toy, doll, smooth plastic skin, clothes, shirt, apron, deformed paws, human hands, blurry, oversaturated

Required Models 4 Models

Disk Space Required: 12.3 GB (4 models · DiT/Base: 6.8 GB · Text Encoder: 5.1 GB · VAE: 335 MB)

models/diffusion_models/ 6.8 GB HuggingFace

flux1-dev-Q4_K_S.gguf Flux.1 Dev GGUF Q4_K_S quantized diffusion backbone

Get

models/text_encoders/ 246 MB HuggingFace

clip_l.safetensors OpenCLIP ViT-L text encoder

Get

models/text_encoders/ 4.9 GB HuggingFace

t5xxl_fp8_e4m3fn.safetensors T5-XXL FP8 language understanding encoder

Get

models/vae/ 335 MB HuggingFace

ae.safetensors Standard 16-channel Flux Autoencoder VAE

Get

Field Notes RTX 4070 Benchmark

Benchmark: RTX 4070 (12GB VRAM): ~22.6 seconds per 768x1024 frame. Peak VRAM 9.8GB at 25 steps.

Photorealistic Realism Tuning: Defeated the default Flux.1 Dev aesthetic over-smoothing and cartoon plushie bias. Stripped anthropomorphic clothing (no pink shirt) and enforced binchotan charcoal under-glow, focused squinting canine eyes against cooking smoke, bubbling tare glaze on negima skewers with charred edges, and realistic optical 35mm film grain.

Frequently Asked Questions FAQ

What GPU and VRAM are required to run Civitai #143442077 Reproduction: Shiba Inu Izakaya Yakitori Chef?

This workflow requires a minimum of 10GB VRAM (recommended 12GB VRAM). Tested and verified on NVIDIA GeForce RTX 4070 (12GB VRAM) at 768x1024 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 quantization format and CLIP encoders should I load for Flux.1 Dev?

This workflow utilizes Flux.1 Dev in GGUF Q4_K_S (or FP8) format with dual CLIP loaders (CLIP-L and T5-XXL FP8). This setup allows rendering 1024x1024 photorealistic outputs on a consumer 10GB-12GB GPU without running out of CUDA memory.