Diffusers代码学习：Textual inversion

[StableDiffusionPipeline] 支持Textual inversion，这种技术使得像 Stable Diffusion 这样的模型能够从几个示例图像中学习新概念。这样可以更好地控制生成的图像，并允许根据特定概念定制模型。

import torch
from diffusers import StableDiffusionPipeline
from diffusers.utils import make_image_grid

pretrained_model_name_or_path = "runwayml/stable-diffusion-v1-5"
repo_id_embeds = "sd-concepts-library/cat-toy"

pipeline = StableDiffusionPipeline.from_pretrained(
    pretrained_model_name_or_path, torch_dtype=torch.float16, use_safetensors=True
).to("cuda")

pipeline.load_textual_inversion(repo_id_embeds)

prompt = "a grafitti in a favela wall with a <cat-toy> on it"

num_samples_per_row = 2
num_rows = 2

all_images = []
for _ in range(num_rows):
    images = pipeline(prompt, num_images_per_prompt=num_samples_per_row, num_inference_steps=50,
                      guidance_scale=7.5).images
    all_images.extend(images)

grid = make_image_grid(all_images, num_rows, num_samples_per_row)

grid.show()