YouTube Thumbnail Kurt Image Generator

Transform your visual content with the YouTube Thumbnail Kurt Image Generator. Ready to experience the power of AI? Start your journey here!

Platform: Replicate

YouTube ThumbnailsImage InpaintingLoRA Fine-tuningAuto-generated Content

151 runs

H100

License Check Required

🚀Function Overview

Generates and edits images for YouTube thumbnails with customizable parameters and LoRA support.

Key Features

Text-to-image conversion using prompts
Image-to-image transformation mode
Inpainting using image masks
Aspect ratio/resolution customization
Adjustable diffusion parameters
LoRA model fine-tuning integration
Fast-generation optimization

Use Cases

•YouTube thumbnail creation
•Social media content personalization
•Branded concept imagery generation
•Marketing material prototyping

⚙️Input Parameters

prompt

string

Prompt for generated image. If you include the `trigger_word` used in the training process you are more likely to activate the trained object, style, or concept in the resulting image.

image

string

Input image for image to image or inpainting mode. If provided, aspect_ratio, width, and height inputs are ignored.

mask

string

Image mask for image inpainting mode. If provided, aspect_ratio, width, and height inputs are ignored.

aspect_ratio

string

Aspect ratio for the generated image. If custom is selected, uses height and width below & will run in bf16 mode

height

integer

Height of generated image. Only works if `aspect_ratio` is set to custom. Will be rounded to nearest multiple of 16. Incompatible with fast generation

width

integer

Width of generated image. Only works if `aspect_ratio` is set to custom. Will be rounded to nearest multiple of 16. Incompatible with fast generation

prompt_strength

number

Prompt strength when using img2img. 1.0 corresponds to full destruction of information in image

model

string

Which model to run inference with. The dev model performs best with around 28 inference steps but the schnell model only needs 4 steps.

num_outputs

integer

Number of outputs to generate

num_inference_steps

integer

Number of denoising steps. More steps can give more detailed images, but take longer.

guidance_scale

number

Guidance scale for the diffusion process. Lower values can give more realistic images. Good values to try are 2, 2.5, 3 and 3.5

seed

integer

Random seed. Set for reproducible generation

output_format

string

Format of the output images

output_quality

integer

Quality when saving the output images, from 0 to 100. 100 is best quality, 0 is lowest quality. Not relevant for .png outputs

disable_safety_checker

boolean

Disable safety checker for generated images.

go_fast

boolean

Run faster predictions with model optimized for speed (currently fp8 quantized); disable to run in original bf16

megapixels

string

Approximate number of megapixels for generated image

lora_scale

number

Determines how strongly the main LoRA should be applied. Sane results between 0 and 1 for base inference. For go_fast we apply a 1.5x multiplier to this value; we've generally seen good performance when scaling the base value by that amount. You may still need to experiment to find the best value for your particular lora.

extra_lora

string

Load LoRA weights. Supports Replicate models in the format <owner>/<username> or <owner>/<username>/<version>, HuggingFace URLs in the format huggingface.co/<owner>/<model-name>, CivitAI URLs in the format civitai.com/models/<id>[/<model-name>], or arbitrary .safetensors URLs from the Internet. For example, 'fofr/flux-pixar-cars'

extra_lora_scale

number

Determines how strongly the extra LoRA should be applied. Sane results between 0 and 1 for base inference. For go_fast we apply a 1.5x multiplier to this value; we've generally seen good performance when scaling the base value by that amount. You may still need to experiment to find the best value for your particular lora.

💡Usage Examples

Example 1

Input Parameters

{
  "model": "dev",
  "prompt": "Generate an image of youtubethumbnailkurt , closed mouth gazing at the camera, pointing at the camera with right hand while showing a smartphone in left hand to the audience. youtubethumbnailkurt is wearing a black long sleeve shirt, covering arms and is positioned on the right side of the frame. The background shows a busy call centre with subtle code elements. The lighting is bright studio lighting with a slight blue accent light from the left. The image has darker colours with high contrast. . Ultra high resolution, 16:9 aspect ratio, professional photography quality, perfect for YouTube thumbnail. Text overlay space on the left side.\n\nAdditional elements:\n- Floating money icons around the smartphone\n- A small laptop on a desk in the lower corner\n- Subtle lens flare from upper right\n\nThe thumbnail conveys the topic: \"hidden revenue in sales processes\"",
  "go_fast": false,
  "lora_scale": 1,
  "megapixels": "1",
  "num_outputs": 1,
  "aspect_ratio": "9:16",
  "output_format": "png",
  "guidance_scale": 3,
  "output_quality": 80,
  "prompt_strength": 0.8,
  "extra_lora_scale": 1,
  "num_inference_steps": 28
}

Output Results

https://replicate.delivery/xezq/j4mDNIypnUbHKd6MH6zfvE6AO7IQkY9XVvKSaC1L4UHqZFUKA/out-0.png

Quick Actions

Use NowView Documentation

Technical Specifications

Hardware Type: H100
Run Count: 151
Commercial Use: Unknown/Restricted
Platform: Replicate

Related Keywords

YouTube thumbnail creationsocial media content personalizationbranded concept imagery generationmarketing material prototypingtext-to-image conversionimage-to-image transformationinpaintingLoRA model fine-tuning

Related Models

Flash-beta.0 Image Generation

A model for generating and editing images through text prompts, image inputs, and inpainting with LoRA adapters.

Gigidiva Gigi

A model for generating and editing images using text prompts and optional input images or masks.

Golestan AI

Imagine your idea in Qajar Era