trainning detail
what base code do you use to train your alsome dreambooth model？diffusers or kohya, and do you set a tragger word such as SKS or something? Could you please give me some guide.
LOFI no need trigger word, just use it like normally SD model
about training, I use the script developed by kohya, mainly using lora to fine-tune the model
If you need prompt guidance, you can go to civitai, there are many examples, from the community or from my own post
https://civitai.com/models/9052/lofi
LOFI no need trigger word, just use it like normally SD model
about training, I use the script developed by kohya, mainly using lora to fine-tune the model
If you need prompt guidance, you can go to civitai, there are many examples, from the community or from my own post
https://civitai.com/models/9052/lofi
thank you.
LOFI no need trigger word, just use it like normally SD model
about training, I use the script developed by kohya, mainly using lora to fine-tune the model
If you need prompt guidance, you can go to civitai, there are many examples, from the community or from my own post
https://civitai.com/models/9052/lofithank you.
Hi, what do you mean to fine-tune base model by lora, does that mean train a specific task lora, then merge lora to base model?
LOFI no need trigger word, just use it like normally SD model
about training, I use the script developed by kohya, mainly using lora to fine-tune the model
If you need prompt guidance, you can go to civitai, there are many examples, from the community or from my own post
https://civitai.com/models/9052/lofithank you.
Hi, what do you mean to fine-tune base model by lora, does that mean train a specific task lora, then merge lora to base model?
yes, you are right, my main fine-tuning is through LORA technology
Secondly, each iteration may produce multiple versions of the base model, and then test the effect of combining different UNet Block layers