ComfyUI.org

Unlock the Magic of Van Gogh: AI-Generated Sea and Sky Masterpieces

CN
ComfyUI.org
2025-06-20 07:57:40

1. Workflow Overview

mc4ioh9lyft7sbqu5h海天图片.png

This workflow generates Van Gogh-style sea and sky images with:

  • Artistic rendering via GhostMix_V2.0 model

  • Bold brushstrokes, starry backgrounds, and warm-cool contrasts

  • Pure text-to-image pipeline (no img2img components)

2. Core Models

  • GhostMix_V2.0: Artistic SD model specialized in color/texture.

  • VAE Decoder: Converts latent data to images (built-in).

3. Key Nodes

Node Name

Function

Installation

CheckpointLoaderSimple

Load model/CLIP/VAE

Built-in

CLIPTextEncode

Encode positive/negative prompts

Built-in

KSampler

Controls generation (sampler/steps)

Built-in

EmptyLatentImage

Sets resolution (824x1336)

Built-in

4. Workflow Steps

  1. Model Loading:

    • Load GhostMix model with CLIP/VAE.

  2. Prompt Processing:

    • Positive: "van gogh style ocean sky, heavy brushstrokes, starry..."

    • Negative: "blurry, lowres, bad proportions..."

  3. Generation:

    • KSampler (Euler, 98 steps) → VAEDecode to final image.

5. Inputs & Outputs

  • Inputs:

    • Resolution: 824×1336 (via EmptyLatentImage)

    • Seed: Fixed or randomized (e.g., 238532686184733)

  • Output: PNG saved to ComfyUI folder.

6. Notes

  1. Model Setup: Download GhostMix_V2.0 from HuggingFace.

  2. Prompt Tips:

    • Include style keywords (e.g., "van gogh").

    • Negative prompts are pre-filled with quality filters.

  3. Performance:

    • ≥8GB VRAM recommended; reduce steps (e.g., 98→50) for faster runs.

previous
Crystal Clear Videos: A Step-by-Step Guide to Dynamic Cutting Animations
next
Unlock Creative Image Fusion: A Step-by-Step Workflow Guide

Recommend