How do I load and continue Flux Lora training from an existing safetensors Lora file?

[a-l-e-x-d-s-9]

I want to continue my training using a safetensors Lora file I have from a previous training session on a server. I only have the safetensors file for the Lora - no backups or other saved files due to the deletion of the previous server.
Is there a way to load pretrained Lora weights from the Flux Lora and proceed with training using new parameters?

[aa956]

Have you tried the method mentioned in the docs?

https://github.com/bghira/SimpleTuner/blob/main/OPTIONS.md

  --init_lora INIT_LORA
                        Specify an existing LoRA safetensors file to
                        initialize the LoRA and continue training or finetune
                        an existing LoRA.

If it's not working for flux LoRA specifically then it's probably worth a bug report?

[a-l-e-x-d-s-9]

@aa956
Yes, I have used it.
It has been added after this question has been asked.
I only tested regular Lora, not sure it's working with Lycoris.