Model Training: LoRa

Action shot of Adam West Batman in fighting pose with buildings and explosions in the background
Action shot of Adam West Batman in fighting pose with buildings and explosions in the background
Medium shot of Adam West Batman inside a retro computer romp clenching fists
Medium shot of Adam West Batman standing on a rooftop with a city in the background
Medium shot of Adam West Batman standing on a rooftop with a city in the background

Project information

  • Category: Model Training

This project focuses on training a LoRA model for 1966 Batman, the iconic character of the series. LoRA, or Low-Rank Adaptation, enables efficient fine-tuning of diffusion models for specific concepts, like characters, with smaller file sizes and faster training times.

I gathered a diverse dataset of Adam West Batman images, curated for quality and relevance. Captions describing pose, expression, and background were created for each image, serving as prompts for the LoRA model. Training was done on Stable Diffusion XL, optimizing settings and hyperparameters for quality.

The project aimed to explore diffusion models, refining data collection, curation, and training methods. Choosing a complex character challenged the model to achieve optimal accuracy.