import gradio as gr
from transformers import BlipProcessor, BlipForConditionalGeneration
from PIL import Image

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

def generate_caption(image):
    inputs = processor(image, return_tensors="pt")
    outputs = model.generate(**inputs)
    description = processor.decode(outputs[0], skip_special_tokens=True)
    return description

# 创建Gradio接口
gradio_app = gr.Interface(
    fn=generate_caption,
    inputs=gr.Image(type="pil"),
    outputs="text",
    title="图片描述生成器",
    description="上传一张图片，生成相应的描述。"
)

if __name__ == "__main__":
    gradio_app.launch()