diffusers-load adapters_pytorch lora weights-CSDN博客

本文链接：https://blog.csdn.net/qq_45692660/article/details/135052239

一、DreamBooth（待补充）

二、Textual inversion（待补充）

三、Lora

在这里插入图片描述

from diffusers import AutoPipelineForText2Image
import torch

pipeline = AutoPipelineForText2Image.from_pretrained("stabilityai/stable-diffusion-xl-base-1.0", torch_dtype=torch.float16).to("cuda")

在这里插入图片描述

pipeline.load_lora_weights("ostris/super-cereal-sdxl-lora", weight_name="cereal_box_sdxl_v1.safetensors")
prompt = "bears, pizza bites"
image = pipeline(prompt).images[0]
image

在这里插入图片描述

from diffusers import AutoPipelineForText2Image
import torch

pipeline = AutoPipelineForText2Image.from_pretrained("stabilityai/stable-diffusion-xl-base-1.0", torch_dtype=torch.float16).to("cuda")
pipeline.unet.load_attn_procs("jbilcke-hf/sdxl-cinematic-1", weight_name="pytorch_lora_weights.safetensors")

# use cnmt in the prompt to trigger the LoRA
prompt = "A cute cnmt eating a slice of pizza, stunning color scheme, masterpiece, illustration"
image = pipeline(prompt).images[0]
image

注意区分load_lora_weights和unet.load_attn_procs的区别，这里我觉得还是推荐用load_lora_weights

在这里插入图片描述

这里说的无法取消多个lora权重的融合，不是很理解，还需要继续去看diffuser的底层代码

from diffusers import DiffusionPipeline
import torch

pipeline = DiffusionPipeline.from_pretrained("stabilityai/stable-diffusion-xl-base-1.0", torch_dtype=torch.float16).to("cuda")
pipeline.load_lora_weights("ostris/ikea-instructions-lora-sdxl", weight_name="ikea_instructions_xl_v1_5.safetensors", adapter_name="ikea")
pipeline.load_lora_weights("ostris/super-cereal-sdxl-lora", weight_name="cereal_box_sdxl_v1.safetensors", adapter_name="cereal")

在这里插入图片描述

prompt = "A cute brown bear eating a slice of pizza, stunning color scheme, masterpiece, illustration"
image = pipeline(prompt, num_inference_steps=30, cross_attention_kwargs={"scale": 1.0}).images[0]
image