1580158
/
kohya_ss
mirror of https://github.com/bmaltais/kohya_ss
1
0
Fork 0
Code Issues Packages Projects Releases Wiki Activity

Update documentation

pull/978/head
bmaltais 2023-06-10 13:52:08 -04:00
parent 524ea25086
commit ffc400d40e
2 changed files with 38 additions and 21 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

53
docs/image_folder_structure.md
View File

@ -1,29 +1,46 @@
# Drambootd, Lora and TI image folder structure
To ensure successful training, Kohya utilizes a specific folder structure that provides the necessary repeat value for the images. It is essential to follow this structure precisely.
To ensure successful training with Kohya, it is crucial to follow a specific folder structure that provides the necessary image repeats. Please adhere to the following structure precisely:
For instance, let's consider training two concepts, each with 30 repeats. We will refer to the first concept as cat and the second concept as dog. While you need at least one concept for training, you can include as many as desired.
The folder structure for these concepts should be as follows:
Folder Structure Example:
```txt
images
c:
|
├── 30_cat
├──images
| |
| ├── image1.jpg
| ├── image1.txt
| ├── image2.png
| └── image2.txt
| ├── 30_cat
| | |
| | ├── image1.jpg
| | ├── image1.txt
| | ├── image2.png
| | └── image2.txt
| |
| └── 30_dog
| |
| ├── image1.jpg
| ├── image1.txt
| ├── image2.png
| └── image2.txt
|
└── 30_dog
|
├── image1.jpg
├── image1.txt
├── image2.png
└── image2.txt
├──regularization
| |
| ├── 1_cat
| | |
| | ├── reg1.jpg
| | ├── reg2.jpg
| |
| └── 1_dog
| |
| ├── reg1.jpg
| ├── reg2.jpg
```
Please note that the same folder structure is required for regularization images as well.
Ensure that you maintain the same structure for the regularization images.
By adhering to this folder structure, you ensure that the images are organized appropriately and that the training process proceeds smoothly.
In the Kohya_ss GUI, follow these steps:
1. Enter the path to the images folder in the Image folder field.
2. Enter the path to the regularisation folder in the Regularisation folder field.
By adhering to this folder structure, you can ensure a smooth and effective training process with Kohya.

6
library/common_gui.py
View File

@ -1288,7 +1288,7 @@ def run_cmd_advanced_training(**kwargs):
def verify_image_folder_pattern(folder_path):
# Check if the folder exists
if not os.path.isdir(folder_path):
log.error(f"The provided path '{folder_path}' is not a valid folder. Please remediate...")
log.error(f"The provided path '{folder_path}' is not a valid folder. Please follow the folder structure documentation found at docs\image_folder_structure.md ...")
return False
# Create a regular expression pattern to match the required sub-folder names
@ -1303,12 +1303,12 @@ def verify_image_folder_pattern(folder_path):
# Check if all sub-folders match the pattern
if len(matching_subfolders) != len(os.listdir(folder_path)):
log.error(f"Not all images folders have proper name patterns in {folder_path}. Please remediate...")
log.error(f"Not all images folders have proper name patterns in {folder_path}. Please follow the folder structure documentation found at docs\image_folder_structure.md ...")
return False
# Check if no sub-folders exist
if not matching_subfolders:
log.error(f"No image folders found in {folder_path}. Please remediate...")
log.error(f"No image folders found in {folder_path}. Please follow the folder structure documentation found at docs\image_folder_structure.md ...")
return False
log.info(f'Valid image folder names found in: {folder_path}')