使用gradio搭建服务(一)

使用参考gradio官方教程

1. 三种基本输入：image，text，checkbox
2. 多个输出，gr.outputs实现

import gradio as gr
import numpy as np
def predict(prompt,image,is_rain):
    print('prompt',type(prompt))
    print('image',type(image))
    print('is_rain',type(is_rain))

    image = np.fliplr(image)

    if prompt:
        output1 = prompt.upper()
        output2 = prompt.lower()
    else:
        output1=0
        output2 = 0
    return output1, output2, image

outputs = [
    gr.outputs.Textbox(label="Uppercase"),
    gr.outputs.Textbox(label="Lowercase"),
    gr.outputs.Image(type = 'numpy',label="Flipped Image")

]

demo = gr.Interface(
    fn=predict,
    inputs=[ "text",'image','checkbox'],
    outputs=outputs,
)
demo.launch()

输出

prompt <class 'str'>
image <class 'numpy.ndarray'>
is_rain <class 'bool'>

图片左边3个框是输入部分，显示的名字为传入参数的名字。右边的三个框是输出部分，名字是output中定义的label

• 如果要同一局域网的其他设备访问服务，设置最后启动的launch(server_name="0.0.0.0")，访问的时候将127.0.0.1更换成运行设备的ip

显示表格

import gradio as gr
import numpy as np

import gradio as gr
import pandas as pd

def display_table(image):
    data = {'Name': ['Alice', 'Bob', 'Charlie'],
            'Age': [25, 30, 35],
            'City': ['New York', 'London', 'Paris']}
    df = pd.DataFrame(data)
    
    return df


# iface.launch()

if __name__ == "__main__":
    # demo.launch()
    demo = gr.Interface(fn=display_table,inputs=['image'], outputs="dataframe")
    demo.launch(server_name="0.0.0.0") ## ip地址+端口号

image
inputs=gr.Image(type="pil")

参考示例

1. https://blog.csdn.net/sinat_26917383/article/details/133950480
2. https://www.gradio.app/guides/image-classification-in-pytorch
3. https://blog.csdn.net/2201_75761617/article/details/132385304
4. 搭建多模态问答机器人https://github.com/airaria/Visual-Chinese-LLaMA-Alpaca/blob/main/scripts/inference/gradio_demo.py