192 lines
6.6 KiB
JSON
192 lines
6.6 KiB
JSON
{
|
|
"_preset_prompts": [
|
|
"🖼️ Tags",
|
|
"🖼️ Simple Description",
|
|
"🖼️ Detailed Description",
|
|
"🖼️ Ultra Detailed Description",
|
|
"🎬 Cinematic Description",
|
|
"🖼️ Detailed Analysis",
|
|
"📹 Video Summary",
|
|
"📖 Short Story",
|
|
"🧩Prompt Refine & Expand"
|
|
],
|
|
"_system_prompts": {
|
|
"🖼️ Tags": "Your task is to generate a clean list of comma-separated tags for a text-to-image AI, based *only* on the visual information in the image. Limit the output to a maximum of 50 unique tags. Strictly describe visual elements like subject, clothing, environment, colors, lighting, and composition. Do not include abstract concepts, interpretations, marketing terms, or technical jargon (e.g., no 'SEO', 'brand-aligned', 'viral potential'). The goal is a concise list of visual descriptors. Avoid repeating tags.",
|
|
"🖼️ Simple Description": "Analyze the image and write a single concise sentence that describes the main subject and setting. Keep it grounded in visible details only.",
|
|
"🖼️ Detailed Description": "Generate a detailed paragraph that combines the subject, actions, environment, lighting, and mood into 2-3 cohesive sentences. Focus on accurate visual details rather than speculation.",
|
|
"🖼️ Ultra Detailed Description": "Produce an extremely rich description touching on appearance, clothing textures, background elements, light quality, shadows, and atmosphere. Aim for an immersive depiction rooted in what the image shows.",
|
|
"🎬 Cinematic Description": "Describe the scene as if capturing a cinematic shot. Cover subject, pose, environment, lighting, mood, and artistic style (photorealistic, painterly, etc.) in one vivid paragraph emphasizing visual impact.",
|
|
"🖼️ Detailed Analysis": "Describe this image in detail, breaking down the subject, attire, accessories, background, and composition into separate sections.",
|
|
"📹 Video Summary": "Summarize the key events and narrative points in this video.",
|
|
"📖 Short Story": "Write a short, imaginative story inspired by this image or video.",
|
|
"🧩Prompt Refine & Expand": "Refine and enhance the following user prompt for creative text-to-image generation. Keep the meaning and keywords, make it more expressive and visually rich. Output **only the improved prompt text itself**, without any reasoning steps, thinking process, or additional commentary."
|
|
},
|
|
"Qwen3-VL-2B-Instruct": {
|
|
"repo_id": "Qwen/Qwen3-VL-2B-Instruct",
|
|
"default": true,
|
|
"quantized": false,
|
|
"vram_requirement": {
|
|
"full": 4.0,
|
|
"8bit": 2.5,
|
|
"4bit": 1.5
|
|
}
|
|
},
|
|
"Qwen3-VL-2B-Thinking": {
|
|
"repo_id": "Qwen/Qwen3-VL-2B-Thinking",
|
|
"default": false,
|
|
"quantized": false,
|
|
"vram_requirement": {
|
|
"full": 4.0,
|
|
"8bit": 2.5,
|
|
"4bit": 1.5
|
|
}
|
|
},
|
|
"Qwen3-VL-2B-Instruct-FP8": {
|
|
"repo_id": "Qwen/Qwen3-VL-2B-Instruct-FP8",
|
|
"default": false,
|
|
"quantized": true,
|
|
"vram_requirement": {
|
|
"full": 2.5
|
|
}
|
|
},
|
|
"Qwen3-VL-2B-Thinking-FP8": {
|
|
"repo_id": "Qwen/Qwen3-VL-2B-Thinking-FP8",
|
|
"default": false,
|
|
"quantized": true,
|
|
"vram_requirement": {
|
|
"full": 2.5
|
|
}
|
|
},
|
|
"Qwen3-VL-4B-Instruct": {
|
|
"repo_id": "Qwen/Qwen3-VL-4B-Instruct",
|
|
"default": true,
|
|
"quantized": false,
|
|
"vram_requirement": {
|
|
"full": 6.0,
|
|
"8bit": 3.5,
|
|
"4bit": 2.0
|
|
}
|
|
},
|
|
"Qwen3-VL-4B-Thinking": {
|
|
"repo_id": "Qwen/Qwen3-VL-4B-Thinking",
|
|
"default": false,
|
|
"quantized": false,
|
|
"vram_requirement": {
|
|
"full": 6.0,
|
|
"8bit": 3.5,
|
|
"4bit": 2.0
|
|
}
|
|
},
|
|
"Qwen3-VL-4B-Instruct-FP8": {
|
|
"repo_id": "Qwen/Qwen3-VL-4B-Instruct-FP8",
|
|
"default": false,
|
|
"quantized": true,
|
|
"vram_requirement": {
|
|
"full": 2.5
|
|
}
|
|
},
|
|
"Qwen3-VL-4B-Thinking-FP8": {
|
|
"repo_id": "Qwen/Qwen3-VL-4B-Thinking-FP8",
|
|
"default": false,
|
|
"quantized": true,
|
|
"vram_requirement": {
|
|
"full": 2.5
|
|
}
|
|
},
|
|
"Qwen3-VL-8B-Instruct": {
|
|
"repo_id": "Qwen/Qwen3-VL-8B-Instruct",
|
|
"default": false,
|
|
"quantized": false,
|
|
"vram_requirement": {
|
|
"full": 12.0,
|
|
"8bit": 7.0,
|
|
"4bit": 4.5
|
|
}
|
|
},
|
|
"Qwen3-VL-8B-Thinking": {
|
|
"repo_id": "Qwen/Qwen3-VL-8B-Thinking",
|
|
"default": false,
|
|
"quantized": false,
|
|
"vram_requirement": {
|
|
"full": 12.0,
|
|
"8bit": 7.0,
|
|
"4bit": 4.5
|
|
}
|
|
},
|
|
"Qwen3-VL-8B-Instruct-FP8": {
|
|
"repo_id": "Qwen/Qwen3-VL-8B-Instruct-FP8",
|
|
"default": false,
|
|
"quantized": true,
|
|
"vram_requirement": {
|
|
"full": 7.5
|
|
}
|
|
},
|
|
"Qwen3-VL-8B-Thinking-FP8": {
|
|
"repo_id": "Qwen/Qwen3-VL-8B-Thinking-FP8",
|
|
"default": false,
|
|
"quantized": true,
|
|
"vram_requirement": {
|
|
"full": 7.5
|
|
}
|
|
},
|
|
"Qwen3-VL-32B-Instruct": {
|
|
"repo_id": "Qwen/Qwen3-VL-32B-Instruct",
|
|
"default": false,
|
|
"quantized": false,
|
|
"vram_requirement": {
|
|
"full": 28.0,
|
|
"8bit": 14.0,
|
|
"4bit": 8.5
|
|
}
|
|
},
|
|
"Qwen3-VL-32B-Thinking": {
|
|
"repo_id": "Qwen/Qwen3-VL-32B-Thinking",
|
|
"default": false,
|
|
"quantized": false,
|
|
"vram_requirement": {
|
|
"full": 28.0,
|
|
"8bit": 14.0,
|
|
"4bit": 8.5
|
|
}
|
|
},
|
|
"Qwen3-VL-32B-Instruct-FP8": {
|
|
"repo_id": "Qwen/Qwen3-VL-32B-Instruct-FP8",
|
|
"default": false,
|
|
"quantized": true,
|
|
"vram_requirement": {
|
|
"full": 24.0
|
|
}
|
|
},
|
|
"Qwen3-VL-32B-Thinking-FP8": {
|
|
"repo_id": "Qwen/Qwen3-VL-32B-Thinking-FP8",
|
|
"default": false,
|
|
"quantized": true,
|
|
"vram_requirement": {
|
|
"full": 24.0
|
|
}
|
|
},
|
|
"Qwen2.5-VL-3B-Instruct": {
|
|
"repo_id": "Qwen/Qwen2.5-VL-3B-Instruct",
|
|
"default": false,
|
|
"quantized": false,
|
|
"vram_requirement": {
|
|
"full": 6.0,
|
|
"8bit": 3.5,
|
|
"4bit": 2.0
|
|
}
|
|
},
|
|
"Qwen2.5-VL-7B-Instruct": {
|
|
"repo_id": "Qwen/Qwen2.5-VL-7B-Instruct",
|
|
"default": false,
|
|
"quantized": false,
|
|
"vram_requirement": {
|
|
"full": 15.0,
|
|
"8bit": 8.5,
|
|
"4bit": 5.0
|
|
}
|
|
}
|
|
}
|
|
|
|
|
|
|