Huggingface Diffusers Dreambooth LoRA Trainer for Flux.1-Dev
About
Cog implementation of the Diffusers Dreambooth LoRA Trainer
Follow me on Twitter/X @lucataco93
How to use
Upload images as a zip file to train a LoRA. For max_train_steps, 100 steps per image is recommended. For resolution, 512 is recommended
If you want to upload your trained LoRA to Huggingface, be sure to add a HF_token that has write access and choose a model location.
There is also a wandb option to track your training run.
An example of an uploaded LoRA can be found here: lucataco/flux-queso
Run your LoRA
Copy the URL from the trained_model.tar from the prediction output, and use the lucataco/flux-dev-lora Explorer model to run your LoRA
An example of a prediction run with a LoRA can be found here
Licensing and commercial use
If you generate images on Replicate with FLUX.1 models and their fine-tunes, then you can use the images commercially.
If you download the weights off Replicate and generate images on your own computer, you can’t use the images commercially.