Magiteker

Author: Frank O’Hanlon

  • Retrieval Based Voice Conversion

    –Github Retrieval Based Voice Conversion is a technique for analyzing speech recordings and training an AI model to emulate vocal patterns. Using this system it is possible to create text to speech models with more natural voices. These models can also be utilized to sing lyrics to songs or recite long texts.

  • TBD Experiment: Multiple Classifications

    Hypothesis: Dreambooth can utilize multiple classifications in regularization to improve model training while maintaining model style and data. Each classification should correspond to a tag that is common throughout the training data.

  • Experiment: High Accuracy Training

    Utilizing Kohya_ss GUI I’m experimenting with various settings sourced from different tutorials in the hopes of gathering a basic configuration that can generate accurate renders while using as little time and VRAM as possible during training. Methodology: 1.) To quickly iterate through configurations I’m using a small dataset of 5 images, all with captions, to…

  • Dreambooth Kohya SS Install

    – Github Above is the repository for Kohya SS Dreambooth, to install simply clone the repository to a separate folder using Git then run the install.bat file. Note: For better training speed ( iterations per second ) using RTX 40xx GPU’s it’s advised to use CUDA 11.8 so that Xformers can properly access the newest…

  • Note: How to turn any model into an inpainting model

    – Reddit Possible merging technique to enable inpainting on various models. May have implications for other types of merges.

  • Resource: Dreambooth Training README

    – Github README describing proper methods for constructing a dataset. According to the README Regularization images allow for prior preservation of a models existing dataset.

  • Resource: Training Stable Diffusion with Dreambooth

    –Blog Post This article describes experiments with different Learning Rates in training models using Dreambooth. FTA: Summary of Initial Results To get good results training Stable Diffusion with Dreambooth, it’s important to tune the learning rate and training steps for your dataset. Faces are harder to train. In our experiments, a learning rate of 2e-6…

  • Stable Diffusion Experiments

    AI module experiments

  • Unity Design Patterns: Optimization

    Unity Design Patterns: Optimization

    Some low level optimizations present at Unite Los Angeles and Unite Europe 2016. Unite Los Angeles 2016 Move configuration data to Scriptable Objects. Utilize the new instantiate call when parenting an object. Don’t needlessly parent objects in production. Batch changes to Transforms in order to save on transform change message calls. Optimize game object transforms…

  • Unity Design Patterns: Scriptable Objects

    Unity Design Patterns: Scriptable Objects

    An interesting use of Scriptable Objects within Unity as demonstrated at Unite 2016. video Data Scriptable Object Data Scriptable Manager   Instanced Scriptable Object