SPARK.Chroma_v1

SPARK.Chroma — Model for Realistic Image Generation

Overview

SPARK.Chroma is a fine-tuned model designed for generating realistic images. Based on the Chroma1-HD model with 8.9B parameters and Apache 2.0 license.

[!WARNING] Warning: This model was trained using test fine-tune configuration — outputs may lack realism and fine detail.

<br> <html> <head> <meta charset="UTF-8"> <style> body { display: flex; justify-content: center; align-items: center; min-height: 100vh; margin: 0; background-color: #1a1a1a; font-family: 'Segoe UI', Roboto, Helvetica, Arial, sans-serif; } .card { background: linear-gradient(135deg, #FF9B1C 0%, #FF9000 100%); border-radius: 15px; padding: 15px; width: 100%; box-shadow: 0 20px 40px rgba(0,0,0,0.4); color: #FFFFFF; border: 4px solid #FFA500; } h2 { color: #FFFFFF; font-size: 1.4rem; margin: 0; letter-spacing: 0.5px; text-transform: uppercase; font-weight: 800; } .grid-container { display: grid; grid-template-columns: repeat(2, 1fr); gap: 12px; margin-bottom: 20px; } .tile-container { display: grid; grid-template-columns: repeat(1, 1fr); gap: 12px; margin-bottom: 20px; } .tile { background-color: #FB8C00; padding: 15px; border-radius: 14px; border-bottom: 3px solid #E65100; transition: transform 0.2s ease, background-color 0.2s ease; } .tile:hover { transform: translateY(-3px); background-color: #FFA000; } .tile strong { display: block; font-size: 1.0rem; color: #FFE0B2; margin-bottom: 5px; text-transform: uppercase; } .tile span { display: block; font-size: 0.95rem; font-weight: 600; line-height: 1.2; } </style> </head> <body> <div class="card"> <div class="grid-container"> <div class="tile"> <strong>Start Date</strong> <span>Feb 25, 2026</span> </div> <div class="tile"> <strong>Last Update</strong> <span>May 3, 2026</span> </div> </div> <div class="tile-container"> <div class="tile"> <strong>Local Equipment</strong> <span>CPU: Intel Core I5-12400; GPU: RTX 4090 (24GB); RAM: 128GB; SSD: M.2 NVMe 1024GB</span> </div> </div> <div class="grid-container"> <div class="tile"> <strong>Dataset Size</strong> <span>~2500 Images [512/1024]</span> </div> <div class="tile"> <strong>Training Time</strong> <span>~ 88h 20m [STEP1]</span> <span>~ 109h 40m [STEP2]</span> </div> </div> </div> </body> </html> <br>

You can support me directly on my <a href="https://boosty.to/sg_161222" style="color: #F15F2C"><b>Boosty</b></a>

<hr>

How to Use

For ComfyUI:

• Place the ae.safetensors in the following path: ComfyUI_windows_portable\ComfyUI\models\vae
• Place t5xxl_fp16.safetensors in the following path: ComfyUI_windows_portable\ComfyUI\models\clip
• Place the SPARK.Chroma_v1_1024.safetensors in the following path: ComfyUI_windows_portable\ComfyUI\models\unet or ComfyUI_windows_portable\ComfyUI\models\diffusion_models
• Load the BasicWorkflow.json into ComfyUI and run

<hr>

Optimal Generation Settings

• Sampling Method and Schedule Type: dpmpp_2m simple
• Sampling Steps: 40
• CFG Scale: 3.5-4.0
• Negative Prompt: low quality, ugly, unfinished, out of focus, deformed, disfigure, blurry, smudged, restricted palette, flat colors, noisy, artifacts, fake, generated

<hr>

Limitations

• This version of the model does not have a variety of styles and is intended for realism.
• There may be issues with anatomy and text.