Finetuning of flux or scratch training

[preethamp0197]

I am new to this field and wanted to know if Is there any code available for training the flux from scratch or even finetuning the existing model. All I see is the dreambooth or Lora finetuning.

[github-actions[bot]]

This issue has been automatically marked as stale because it has not had recent activity. If you think this still needs to be addressed please comment on this thread.

Please note that issues that do not follow the contributing guidelines are likely to be ignored.

[asomoza]

Flux is a non commercial open weights model, this means we don't have the training code or the dataset to train it from scratch. People have tried to fine tune it but it's hard because the model that was released it's a distilled model from the teacher (pro) so it's breaks really easy when trying to fine tune it.

Since the scope of this library is to provide some easy to understand and learning code, this is out of scope for the moment. You will need to look into the specialized libraries for training models.

[github-actions[bot]]

This issue has been automatically marked as stale because it has not had recent activity. If you think this still needs to be addressed please comment on this thread.

Please note that issues that do not follow the contributing guidelines are likely to be ignored.