app.py

import streamlit as st
from PIL import Image
from transformers import pipeline
from gtts import gTTS

st.set_page_config(page_title="Your Image to Audio Story", page_icon="🦜")

def generate_caption(image_file):
    image = Image.open(image_file)
    caption_generator = pipeline("image-to-text", model="Salesforce/blip-image-captioning-base")
    caption_results = caption_generator(image)
    caption = caption_results[0]['generated_text']
    return caption

def generate_story(caption):
    story_generator = pipeline("text-generation", model="Qwen/QwQ-32B")
    messages = [
        {
            "role": "user",
            "content": f"Please based on following image caption: '{caption}', generate a complete fairy tale story for children with at least 100 words and max 300 words"
        }
    ]
    result = story_generator(messages, max_length=300, num_return_sequences=1)
    story = result[0]['generated_text']
    return story

@st.cache_resource
# def load_image_generator():
#     device = "cuda" if torch.cuda.is_available() else "cpu"
#     torch_dtype = torch.float16 if device == "cuda" else torch.float32
#     pipe = DiffusionPipeline.from_pretrained("stable-diffusion-v1-5/stable-diffusion-v1-5")
#     pipe = pipe.to(device)
#     return pipe

# def generate_illustration(prompt):
#     pipe = load_image_generator()
#     image_result = pipe(prompt)
#     generated_image = image_result.images[0]
#     return generated_image

def text_to_speech(text, output_file="output.mp3"):
    tts = gTTS(text=text, lang="en")
    tts.save(output_file)
    return output_file

def main():
    st.markdown("<h1 style='text-align: center;'>Your Image to Audio Story 🦜</h1>", unsafe_allow_html=True)
    st.write("Upload an image below and we will generate an engaging story from the picture, then convert the story into an audio playback!")
    
    uploaded_file = st.file_uploader("Select Image", type=["png", "jpg", "jpeg"])
    
    if uploaded_file is not None:
        image = Image.open(uploaded_file)
        st.image(image, caption="Uploaded image", use_container_width=True)
        
        with st.spinner("Image caption being generated..."):
            caption = generate_caption(uploaded_file)
        st.write("**Image Caption:**", caption)
        
        with st.spinner("Generating story..."):
            story = generate_story(caption)
        st.write("**Story:**")
        st.write(story)
        
        # with st.spinner("Generating illustration..."):
        #     illustration = generate_illustration(story[:200])
        # st.write("### Story Illustrations：")
        # st.image(illustration, caption="Story Illustrations", use_container_width=True)
        
        with st.spinner("Converting to voice...."):
            audio_file = text_to_speech(story)
        st.audio(audio_file, format="audio/mp3")

if __name__ == "__main__":
    main()