mathdiagram 1.0.1

Generic mathematical diagram generation library for any topic

ImageGen AI

A Python library for AI-powered image generation using Stable Diffusion models.

Features

• Easy-to-use API for image generation
• Support for batch generation
• Memory management and optimization
• Metadata support for generated images
• Utility functions for image processing
• Configurable settings

Installation

From PyPI (when published)

pip install imagegen-ai

From source

git clone <repository-url>
cd imagegen-ai
pip install -e .

Quick Start

from imagegen import ImageGenerator, save_image

# Initialize generator
generator = ImageGenerator()

# Generate an image
image = generator.generate(
    prompt="A beautiful sunset over mountains",
    width=512,
    height=512,
    seed=42
)

# Save the image
save_image(image, "sunset.png")

Advanced Usage

Batch Generation

prompts = [
    "A futuristic city",
    "A peaceful forest",
    "A vintage car"
]

images = generator.generate_batch(prompts)

Custom Configuration

from imagegen import Config

config = Config(
    model_id="runwayml/stable-diffusion-v1-5",
    device="cuda",
    default_width=768,
    default_height=768
)

generator = ImageGenerator(
    model_id=config.model_id,
    device=config.device
)

Creating Image Grids

from imagegen.utils import create_grid

# Generate multiple images
images = generator.generate_batch(prompts)

# Create a grid
grid = create_grid(images, grid_size=(2, 2))
save_image(grid, "image_grid.png")

API Reference

ImageGenerator

Main class for image generation.

Methods

• generate(prompt, **kwargs) - Generate a single image
• generate_batch(prompts, **kwargs) - Generate multiple images
• change_model(model_id) - Switch to a different model
• get_memory_usage() - Get current memory usage
• clear_memory() - Clear GPU memory cache

Utility Functions

• save_image(image, filename, directory, metadata) - Save a single image
• save_images(images, base_filename, directory, metadata) - Save multiple images
• create_grid(images, grid_size, spacing) - Create an image grid
• batch_resize(images, size, method) - Resize multiple images

Requirements

• Python 3.8+
• PyTorch 1.13.0+
• CUDA-capable GPU (recommended)

License

MIT License - see LICENSE file for details.