python用魔塔大模型生成国画

模型地址 https://www.modelscope.cn/models/langboat/Guohua-Diffusion/summary

新建一个main.py

from modelscope.utils.constant import Tasks
from modelscope.pipelines import pipeline
import cv2

pipe = pipeline(task=Tasks.text_to_image_synthesis, 
                model='langboat/Guohua-Diffusion',
                model_revision='v1.0')

prompt = 'The Godfather poster in guohua style'
output = pipe({'text': prompt})
cv2.imwrite('result.png', output['output_imgs'][0])

安装依赖

modelscope
torch
transformers
diffusers 

运行会自动去魔塔下载模型文件，结果大概为

默认模型会下载到~/.cache/modelscope/hub中，如果需要修改下载目录，可以手动指定环境变量：MODELSCOPE_CACHE，modelscope会将模型和数据集下载到该环境变量指定的目录中。

如果是本地下载好的大模型

from modelscope.utils.constant import Tasks
from modelscope.pipelines import pipeline
import cv2
# 目录路径,一直到模型
download_path = f'C:/Users/PC/.cache/modelscope/hub/langboat/Guohua-Diffusion'

pipe = pipeline(task=Tasks.text_to_image_synthesis, 
                model=download_path,
                model_revision='v1.0')

prompt = 'chinese girl in guohua style'
output = pipe({'text': prompt})
cv2.imwrite('result.png', output['output_imgs'][0])