https://github.com/invoke-ai/invoke-training
https://github.com/invoke-ai/invoke-training
Last synced: about 1 year ago
JSON representation
- Host: GitHub
- URL: https://github.com/invoke-ai/invoke-training
- Owner: invoke-ai
- License: apache-2.0
- Created: 2023-07-31T13:23:51.000Z (almost 3 years ago)
- Default Branch: main
- Last Pushed: 2024-06-09T16:35:30.000Z (almost 2 years ago)
- Last Synced: 2024-07-22T19:23:25.244Z (almost 2 years ago)
- Language: Python
- Homepage: https://invoke-ai.github.io/invoke-training/
- Size: 59 MB
- Stars: 86
- Watchers: 8
- Forks: 11
- Open Issues: 8
-
Metadata Files:
- Readme: README.md
- Contributing: docs/contributing/development_environment.md
- License: LICENSE
Awesome Lists containing this project
README
# invoke-training
A library for training custom Stable Diffusion models (fine-tuning, LoRA training, textual inversion, etc.) that can be used in [InvokeAI](https://github.com/invoke-ai/InvokeAI).
> [!WARNING] > `invoke-training` is still under active development, and breaking changes are likely. Full backwards compatibility will not be guaranteed until v1.0.0.
> In the meantime, I recommend pinning to a specific commit hash.
## Documentation
## Training Modes
- Stable Diffusion
- LoRA
- DreamBooth LoRA
- Textual Inversion
- Stable Diffusion XL
- Full finetuning
- LoRA
- DreamBooth LoRA
- Textual Inversion
- LoRA and Textual Inversion
More training modes coming soon!
## Installation
See the [Installation](https://invoke-ai.github.io/invoke-training/get-started/installation/) section of the documentation.
## Quick Start
`invoke-training` pipelines can be configured and launched from either the CLI or the GUI.
### CLI
Run training via the CLI with type-checked YAML configuration files for maximum control:
```bash
invoke-train --cfg-file src/invoke_training/sample_configs/sdxl_textual_inversion_gnome_1x24gb.yaml
```
### GUI
Run training via the GUI for a simpler starting point.
```bash
invoke-train-ui
# Or, you can optionally override the default host and port:
invoke-train-ui --host 0.0.0.0 --port 1234
```
## Features
Training progress can be monitored with [Tensorboard](https://www.tensorflow.org/tensorboard):
_Validation images in the Tensorboard UI._
All trained models are compatible with InvokeAI:
_Example image generated with the prompt "A cute yoda pokemon creature." and a trained Pokemon LoRA._
## Contributing
Contributors are welcome. For developer guidance, see the [Contributing](https://invoke-ai.github.io/invoke-training/contributing/development_environment/) section of the documentation.