app.py

import gradio as gr
from transformers import pipeline

# 创建 pipeline，使用指定模型进行图像到文本的转换
pipe = pipeline("image-to-text", model="microsoft/trocr-base-handwritten")

def recognize_text_from_image(input_image):
    # 使用 pipeline 处理上传的图片
    output = pipe(input_image)
    # 返回识别出的文本
    return output[0]['generated_text']

# 创建 Gradio 接口
iface = gr.Interface(
    fn=recognize_text_from_image,  # 指定处理函数
    inputs=gr.Image(tool="editor", type="pil", label="Upload your image here or use your camera"),  # 设置输入为图片
    outputs="text",  # 输出类型为文本
    title="Image to Text Recognition",  # 接口标题
    description="Upload an image and the text in the image will be recognized."  # 接口描述
)

# 启动 Gradio 应用
iface.launch()