Mirrors
/
EveryDream2trainer
mirror of https://github.com/victorchall/EveryDream2trainer.git
1
0
Fork 0
Code Issues Packages Projects Releases Wiki Activity

validation docs

This commit is contained in:
Victor Hall 2023-02-08 16:11:09 -05:00
parent 2353c4c16a
commit 0eb5a0c0e5
2 changed files with 7 additions and 1 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

2
README.md
View File

@ -41,6 +41,8 @@ Make sure to check out the [tools repo](https://github.com/victorchall/EveryDrea
[Data Balancing](doc/BALANCING.md) - Includes my small treatise on model preservation with ground truth data
[Validation](doc/VALIDATION.md) - Use a validation split on your data to see when you are overfitting and tune hyperparameters
## Cloud
[Free tier Google Colab notebook](https://colab.research.google.com/github/victorchall/EveryDream2trainer/blob/main/Train_Colab.ipynb)

6
doc/VALIDATION.md
View File

@ -1,9 +1,13 @@
# Validation
Validation allows you to split data for evaluating your training progress.
*This documentation is incomplete. Please feel free to contribute to it.*
Validation allows you to use a split of your data for evaluating your training progress.
When training a specific class, setting aside a portion of the data for validation will allow you to see trend lines you cannot see when purely looking at loss of the training itself.
While loss on your training data should trend downward, if you set aside a validation set, you can see when your validation loss starts to trend upward. This is a sign that you are overfitting. You can then adjust your hyperparameters to reduce overfitting, such as reducing LR or reducing training epochs.
## How to use validation
The `validation_config` option is a pointer to a JSON config file with settings for use in validation. There is a default validation file `validation_default.json` in the repo root, but it is not used unless you specify it.