Notebooks to finetune several models that can run on Google Colab.
For the dataset sizes, usually 5 ~ 20 images are enough to train LoRAs.
| Category | Model Type | Notebook | Open in Colab |
|---|---|---|---|
| Flux.1-dev | LoRA |  |
 |
| Hunyuan Video | LoRA | |
|
| SDXL | LoRA | |
|
| LTX Video | LoRA | |
|
| Diffusers Converter (Util) | - | |
|
- Flux LoRA Training : https://github.com/ostris/ai-toolkit
- Hunyuan Video & LTX Video LoRA Training : https://github.com/tdrussell/diffusion-pipe
- Many training scripts are from the diffusers 🤗: https://github.com/huggingface/diffusers