Model Training: LoRa

Batgirl in a toxic waste factory
The image showcases a woman dressed in a purple Batman costume, complete with a bat emblem on her chest and a cape. She appears to be in an urban setting, with buildings in the background. The woman has a surprised or shocked expression on her face. The image seems to be from a vintage or retro style, possibly from the 1960s or 1970s, given the color palette and quality.
Batgir in a fighting stance inside an evil villan's lair
Batgir in a fighting stance inside an evil villan's lair
The image showcases a woman dressed in a purple Batman costume, complete with a mask, cape, and utility belt. She is striking a dynamic pose with one fist raised, as if ready for action. The backdrop features a city skyline with tall buildings, suggesting an urban setting. The image appears to be from a vintage or retro style, possibly from the 1960s or 1970s, given the color palette and quality. The woman's expression is determined, and she exudes a sense of confidence.
The image showcases a woman dressed in a superhero costume, specifically Batgirl, from the Batman series. She is flexing her muscles, showcasing her strength and confidence. The setting appears to be a rooftop or high vantage point overlooking a cityscape. The attire is predominantly purple with a yellow cape. The woman's expression is determined and focused. The image is in color, and the style is vintage, reminiscent of the 1960s or 1970s comic book artwork.
Batgirl working in a retro computer room.

Project information

  • Category: Model Training

The aim of this project is to train a LoRA model for Yvonne Craig Batgirl from the 1966 Batman Series. I chose this character because it is very complex and my goal is to keep practicing with subjects that push the boundaries of the training even further. LoRA, or Low-Rank Adaptation, is a technique that allows efficient fine-tuning of diffusion models for specific concepts, such as characters, with smaller file sizes and faster training times.

I collected a diverse dataset of Batgirl, capturing the most varied range of poses and shots. I wanted to train a high-quality model, but it was a difficult project because the source images were blurry. I had to restore the faces one by one without losing the main features of Yvonne Craig. Some of her facial resemblance was lost in the restoration process, but the costume came out amazing. I think this is the best I can achieve given the quality of the source material. I created captions for each image, describing the pose, expression, and background, to serve as prompts for the LoRA model. I trained the model on Stable Diffusion XL, optimizing the settings and hyperparameters for quality.