---
title: Synthetic Data Gen
emoji: 👁
colorFrom: indigo
colorTo: yellow
sdk: gradio
sdk_version: 5.9.0
app_file: app.py
pinned: false
short_description: Synthetic data generation for training using diffusion model
---

Check out the configuration reference at https://huggingface.co/docs/hub/spaces-config-reference


Features:

Stable Diffusion Integration: Leverages the StableDiffusionImg2ImgPipeline for versatile image generation.

Class Misclassification Analysis: Automatically identifies misclassified class pairs for focused image generation.

Advanced Prompt Interpolation: Utilizes spherical linear interpolation (slerp) to generate smooth transitions between text embeddings for more natural image outputs.

Structural Similarity Scoring: Calculates SSIM scores to measure the quality of generated images compared to input references.

Comprehensive Metadata Management: Captures extensive details about the generation process for easy reproducibility and debugging.

Dynamic Directory Management: Automatically creates and organizes directories for class pairs and metadata files.

Highly Configurable: Supports advanced optimization options like torch.compile and DeepCache for performance improvements.