How to use it with diffusers?
#5
by
Mousey
- opened
Hi, how exactly can one use this lora with diffusers? I got the following error when building an image:
OSError: logo-wizard/logo-diffusion-checkpoint does not appear to have a file named pytorch_lora_weights.bin.
Hi! You need to use StableDiffusionPipeline because it's not a lora itself but the complete checkpoint. There is an example:
from diffusers import StableDiffusionPipeline
import torch
pipe = StableDiffusionPipeline.from_pretrained("logo-wizard/logo-diffusion-checkpoint", torch_dtype=torch.float16).to("cuda");
negative_prompt = "low quality, worst quality, bad composition, extra digit, fewer digits, text, inscription, watermark, label, asymmetric" # for example
prompt = "a logo of flower store, flowers, gradient colors, modern, minimalism, vector art, 2d, best quality, centered" # for example
image = pipe(prompt=prompt, negative_prompt=negative_prompt, num_inference_steps=30, guidance_scale=7.5, height=768, width=768).images[0]
eewwann
changed discussion status to
closed
Hi, thanks a lot. I thought this was a lora but as it seems one can load it directly as a model. I will retry now, looks awesome, congratulations on the great work.