Crandi

Folder Structure Hollowstrawberry Kohya Colab Deepwiki

Leo Migdal
-
folder structure hollowstrawberry kohya colab deepwiki

This document explains the directory organization system used across all kohya-colab notebooks. The repository supports two distinct folder organization modes that affect where datasets, outputs, configurations, and logs are stored in Google Drive. Understanding these conventions is essential for proper dataset preparation and training workflow. For information about dataset configuration (multi-folder datasets, repeats, regularization), see Dataset Configuration. For information about output file naming and epoch management, see Training Configuration. The kohya-colab system provides two mutually exclusive folder organization strategies that users must choose when starting any notebook.

This choice affects the entire directory hierarchy and must remain consistent across Dataset Maker and LoRA Trainer notebooks for the same project. The mode is selected via a dropdown parameter in every notebook's main cell and is evaluated using a simple string matching pattern: Sources: Lora_Trainer_XL.ipynb103 Lora_Trainer_XL.ipynb314-325 Lora_Trainer.ipynb99 Lora_Trainer.ipynb215-226 Dataset_Maker.ipynb68 Dataset_Maker.ipynb84-93 This document describes the TOML-based configuration system used by all trainer notebooks to define training parameters and dataset specifications. The configuration system generates two primary files that control the training process: training_config.toml and dataset_config.toml. For information about specific optimizer configurations, see Optimizer Configuration.

For details on dataset folder structures and validation, see Dataset Validation. The configuration system follows a two-stage generation process where user-defined parameters in the notebook cells are transformed into structured TOML files that are consumed by the underlying kohya-ss training scripts. Sources: Lora_Trainer_XL.ipynb447-570 Lora_Trainer.ipynb361-471 All trainers store configuration files in a project-specific folder within Google Drive, with the structure determined by the selected folder_structure option. The Dataset Preparation system provides an end-to-end workflow for acquiring, curating, and annotating image datasets for LoRA training. This document covers the overall architecture and workflow of the Dataset Maker notebooks.

For specific details on individual phases, see Image Acquisition, Duplicate Detection and Curation, Image Tagging and Captioning, and Tag Management. For information about using prepared datasets in training, see LoRA Training. The Dataset Maker notebooks (Dataset_Maker.ipynb Spanish_Dataset_Maker.ipynb) automate the process of creating high-quality datasets for training LoRA models. The system handles: The output is a collection of images with corresponding text files (captions/tags) ready for consumption by the training notebooks. Sources: Dataset_Maker.ipynb1-100 Spanish_Dataset_Maker.ipynb1-100

The Dataset Maker follows a sequential cell execution model where each step (step1_installed_flag, step2_installed_flag, etc.) must complete before subsequent steps can run. This gating mechanism prevents users from executing steps out of order. This page covers advanced usage patterns and customization options for power users who want to extend beyond the default training configuration. These features enable complex dataset structures, professional experiment tracking, and incremental training workflows. For basic training configuration, see LoRA Training. For standard optimizer and scheduler settings, see Optimizer Configuration and Learning Rate Schedulers.

For base configuration file structure, see Configuration System. The custom dataset feature allows you to define complex multi-folder dataset structures with per-folder configuration. This enables mixing different image sets with different repeat counts, regularization folders, and per-subset processing parameters within a single training session. Sources: Lora_Trainer_XL.ipynb786-826 Lora_Trainer.ipynb601-641 Both trainer notebooks expose a custom_dataset variable that accepts a TOML-formatted string defining dataset structure. When set to a non-None value, this overrides the default single-folder dataset configuration derived from project_name.

This page provides a quick start guide for new users of the kohya-colab repository. It covers the prerequisites, initial setup steps, and the basic workflow for preparing a dataset and running your first LoRA training session. For detailed information about dataset preparation techniques, see Dataset Preparation. For comprehensive training configuration options, see LoRA Training. Before beginning, ensure you have the following: The repository provides multiple notebook entry points hosted on Google Colab.

Each notebook can be opened directly in your browser: Alternatively, you can access notebooks directly via Colab URLs formatted as: All notebooks require Google Drive access for persistent storage of datasets, configurations, and trained models. The mounting process is automatic when you run the first cell of any notebook. Accessible Google Colab notebooks for Stable Diffusion Lora training, based on the work of kohya-ss and Linaqruf. If you need support I now have a public Discord server

There was an error while loading. Please reload this page. There was an error while loading. Please reload this page. Not really a bug but couldn't find a Q&A section. Can you please explain more clearly how to specify a regularization folder?

If my folder structure in the google drive is "my drive => loras => sacbf", in which I have a "dataset" folder and a "reg_images" folder, what and how do I need to specify...

People Also Search

This Document Explains The Directory Organization System Used Across All

This document explains the directory organization system used across all kohya-colab notebooks. The repository supports two distinct folder organization modes that affect where datasets, outputs, configurations, and logs are stored in Google Drive. Understanding these conventions is essential for proper dataset preparation and training workflow. For information about dataset configuration (multi-f...

This Choice Affects The Entire Directory Hierarchy And Must Remain

This choice affects the entire directory hierarchy and must remain consistent across Dataset Maker and LoRA Trainer notebooks for the same project. The mode is selected via a dropdown parameter in every notebook's main cell and is evaluated using a simple string matching pattern: Sources: Lora_Trainer_XL.ipynb103 Lora_Trainer_XL.ipynb314-325 Lora_Trainer.ipynb99 Lora_Trainer.ipynb215-226 Dataset_M...

For Details On Dataset Folder Structures And Validation, See Dataset

For details on dataset folder structures and validation, see Dataset Validation. The configuration system follows a two-stage generation process where user-defined parameters in the notebook cells are transformed into structured TOML files that are consumed by the underlying kohya-ss training scripts. Sources: Lora_Trainer_XL.ipynb447-570 Lora_Trainer.ipynb361-471 All trainers store configuration ...

For Specific Details On Individual Phases, See Image Acquisition, Duplicate

For specific details on individual phases, see Image Acquisition, Duplicate Detection and Curation, Image Tagging and Captioning, and Tag Management. For information about using prepared datasets in training, see LoRA Training. The Dataset Maker notebooks (Dataset_Maker.ipynb Spanish_Dataset_Maker.ipynb) automate the process of creating high-quality datasets for training LoRA models. The system ha...

The Dataset Maker Follows A Sequential Cell Execution Model Where

The Dataset Maker follows a sequential cell execution model where each step (step1_installed_flag, step2_installed_flag, etc.) must complete before subsequent steps can run. This gating mechanism prevents users from executing steps out of order. This page covers advanced usage patterns and customization options for power users who want to extend beyond the default training configuration. These fea...