morlako/rhinoplasty_lora

--- license: cc-by-nc-4.0 task_categories: - image-to-image tags: - medical - rhinoplasty - stable-diffusion - lora - before-after size_categories: - n<1K --- # 🏥 Rhinoplasty Before/After Training Dataset Complete training package for fine-tuning Stable Diffusion with LoRA on rhinoplasty before/after images. ## 📦 What's Inside This ZIP contains everything needed to train a LoRA model: - **127 training pairs** (254 images total) - High quality rhinoplasty results - **Train/validation split** (80/20) - Pre-organized dataset structure - **Training script**: `train_lora_sd.py` - Ready-to-use Python script - **Requirements**: `requirements_lora.txt` - All dependencies listed - **Quick start guide**: `LORA_QUICKSTART.md` - Step-by-step instructions - **RunPod automation**: `RUNPOD_LORA_START.sh` - One-command setup ## 🚀 Quick Start ### Option 1: RunPod (Recommended) ```bash # 1. Download and extract wget https://huggingface.co/datasets/morlako/rhinoplasty-lora-training/resolve/main/rhinoplasty_lora_training.zip unzip rhinoplasty_lora_training.zip cd rhinoplasty_lora_training # 2. Run automated setup bash RUNPOD_LORA_START.sh # 3. Wait 2-4 hours - Done!This is a training dataset of before and after rhinoplasty images, containing 127 pairs of high-quality images (totaling 254 images) for fine-tuning the Stable Diffusion model with LoRA. The dataset is already split into training and validation sets (80/20) and includes a training script, a dependency list, and a quick start guide.

搜集汇总

背景与挑战

背景概述

该数据集为鼻整形前后对比图像数据集，用于微调Stable Diffusion模型的LoRA权重。包含127对高分辨率图像（共254张），并提供了训练脚本、依赖列表和快速启动指南，适合医学图像到图像生成任务。数据集采用非商业许可（cc-by-nc-4.0），总大小约9.76 MB。