File size: 2,298 Bytes
1fa4382
38459f0
1fa4382
 
 
 
 
 
 
 
 
 
 
e3f6286
 
 
 
 
 
 
87a5e54
 
 
 
 
 
 
1fa4382
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
858d64d
1fa4382
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
# import part
from transformers import pipeline
import streamlit as st

# function part
# img2text
def img2text(url):
    image_to_text_model = pipeline("image-to-text", model="nlpconnect/vit-gpt2-image-captioning")
    text = image_to_text_model(url)[0]["generated_text"]
    return text

# text2story
def text2story(text):
    story_prompt = (
        f"You are an expert in storytelling for children ages 3-10."
        f"Based on the image description below:{text}\n"
        f"Please create a funny, heartwarming, imaginative fairy tale of no less than 50 words"
        f"Output only the body of the story, do not include other content"
    )
    
    story_text = pipeline("text-generation", model="distilbert/distilgpt2",
                              max_length=200, 
                              do_sample=True, 
                              top_p=0.95, 
                              temperature=0.7)
    generation = story_text(story_prompt)
    story_text = generation[0]['generated_text']
    return story_text

# text2audio
def text2audio(story_text):
    audio_data = pipeline("text-to-audio", model="facebook/musicgen-medium")
    return audio_data

# main part
st.set_page_config(page_title="Your Image to Audio Story",
                   page_icon="🦜")
st.header("Turn Your Image to Audio Story")

# Upload image here
uploaded_file = st.file_uploader("Select an Image...")

if uploaded_file is not None:
    print(uploaded_file)
    bytes_data = uploaded_file.getvalue()
    with open(uploaded_file.name, "wb") as file:
        file.write(bytes_data)
    st.image(uploaded_file, caption="Uploaded Image",
             use_container_width=True)

    #Stage 1: Image to Text
    st.text('Processing img2text...')
    scenario = img2text(uploaded_file.name)
    st.write(scenario)

    #Stage 2: Text to Story
    st.text('Generating a story...')
    story = text2story(scenario)
    st.write(story)

    #Stage 3: Story to Audio data
    st.text('Generating audio data...')
    audio_data =text2audio(story)

    # Play button
    if st.button("Play Audio"):
        st.audio(audio_data['audio'],
                    format="audio/wav",
                    start_time=0,
                    sample_rate = audio_data['sampling_rate'])
        st.audio("kids_playing_audio.wav")