import streamlit as st
from transformers import BlipProcessor, BlipForConditionalGeneration
from PIL import Image

# 加载BLIP模型和处理器
processor = BlipProcessor.from_pretrained("Salesforce/blip-image-captioning-large")
model = BlipForConditionalGeneration.from_pretrained("Salesforce/blip-image-captioning-large")

st.title("图像描述生成器")
st.write("使用摄像头拍照并生成图像的描述。")

# 使用Streamlit的camera_input来获取用户摄像头输入
image_data = st.camera_input("请使用摄像头拍照")

if image_data is not None:
    # 将图像数据转换为PIL图像
    image = Image.open(image_data)
    
    # 显示拍摄的图像
    st.image(image, caption="拍摄的图像", use_column_width=True)

    # 生成图像描述
    inputs = processor(image, return_tensors="pt")
    out = model.generate(**inputs)
    caption = processor.decode(out[0], skip_special_tokens=True)
    
    st.write(f"图像描述: {caption}")