gradio代码案例+效果图片

直接上代码：

import gradio as gr

import numpy as np
import torch
from PIL import Image
from ram.models import ram_plus
from ram import inference_ram as inference
from ram import get_transform
import time

# 加载模型
m_start = time.time()
device = torch.device('cuda' if torch.cuda.is_available() else 'cpu')
transform = get_transform(image_size=384)
model = ram_plus(pretrained='pretrained/ram_plus_swin_large_14m.pth',
                 image_size=384,
                 vit='swin_l')
model.eval()
model = model.to(device)
m_end = time.time()
print('model load time:{}'.format(m_end-m_start))

# 用于生成随机颜色的函数
def random_color():
    return "#{:06x}".format(np.random.randint(0, 0xFFFFFF))

# Gradio的图像标签回调函数
def image_tagging(input_image):
    # 在这里添加你的图像标签生成代码
    # 这里使用一个示例，你可以替换为你的模型输出
    start = time.time()
    image = transform(input_image).unsqueeze(0).to(device)
    res = inference(image, model)
    labels = res[1].split(' | ')
    end = time.time()
    print('infer time:{}'.format(end-start))
    # labels = ["标签1", "标签2", "标签3"]
    # 生成HTML代码，使用Flexbox布局自适应排列
    label_html = '<div style="display: flex; flex-wrap: wrap;">'
    for label in labels:
        color = random_color()
        label_html += f'<div style="background-color: {color}; color: white; border-radius: 15px; padding: 10px; margin: 5px;">{label}</div>'

    label_html += '</div>'

    description_html = '<p style="font-size: 16px; color: #333;">这里是图片标签的输出结果</p>'

    return description_html+label_html
# Gradio界面配置
iface = gr.Interface(
    fn=image_tagging,
    inputs=gr.Image(type="pil", label="点击上传图片"),
    outputs=gr.HTML(label="此图片包含的标签为"),
    live=False,
    title="智能图像标签体验网页",
    description = "上传图片以后，点击submit按钮，即可提取图片中的各种标签，赶紧试试吧！"

)

# 启动Gradio界面
iface.launch(server_name="0.0.0.0",server_port=8089)

展示界面：