First of all. Thank you for your effort.
I have some question when loading lora weight in sd.
- Is it the right way that putting sd model in this way? I mean it doesn’t have to be converted with like “convert_models_diffuser_to_diffusers.py” in diffusers?
model_base = "/hdd/jinnnn/anylora.safetensors"
pipeline = StableDiffusionImg2ImgPipeline.from_single_file(model_base, torch_dtype=torch.float16, safety_checker=None, requires_safety_checker=False).to('cuda')
Then how can I adjust the weight of the lora weight like
Is there any option that put any argument like “lora_weight” in “load_lora_weights” function?
How can I use the trigger word in prompt? For example if my trigger word is ‘iom’, just putting ‘iom’ in my prompt is applied in generation?
In connection with question 3, can I adjust the lora applied weight with text prompt?
load_lora_weightsis it automatically setting the rank and alpha that was used in trainig in webui?