Commit
·
d1b5c08
1
Parent(s):
1ff7568
Create app.py
Browse files
app.py
ADDED
@@ -0,0 +1,61 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
from transformers import pipeline
|
2 |
+
import streamlit as st
|
3 |
+
import os
|
4 |
+
|
5 |
+
|
6 |
+
|
7 |
+
# img2text
|
8 |
+
def img_to_text(url):
|
9 |
+
image_to_text = pipeline("image-to-text", model="Salesforce/blip-image-captioning-large")
|
10 |
+
|
11 |
+
text = image_to_text(url)[0]["generated_text"]
|
12 |
+
return text
|
13 |
+
|
14 |
+
|
15 |
+
# llm
|
16 |
+
def generate_story(text):
|
17 |
+
generator = pipeline("text-generation", model="distilgpt2")
|
18 |
+
|
19 |
+
result = generator(text, max_length=20, num_return_sequences=1)
|
20 |
+
return result[0]['generated_text']
|
21 |
+
|
22 |
+
|
23 |
+
#
|
24 |
+
# text-to-speech
|
25 |
+
def text_to_speech(text):
|
26 |
+
import requests
|
27 |
+
|
28 |
+
API_URL = "https://api-inference.huggingface.co/models/espnet/kan-bayashi_ljspeech_vits"
|
29 |
+
headers = {"Authorization": f"Bearer {os.environ.get('HUGGINGFACE_API_TOKEN')}"}
|
30 |
+
payload = {
|
31 |
+
"inputs": text
|
32 |
+
}
|
33 |
+
|
34 |
+
response = requests.post(API_URL, headers=headers, json=payload)
|
35 |
+
response.raise_for_status()
|
36 |
+
with open('audio.flac', 'wb') as file:
|
37 |
+
file.write(response.content)
|
38 |
+
|
39 |
+
|
40 |
+
|
41 |
+
def main():
|
42 |
+
st.set_page_config(page_title="img to audio story")
|
43 |
+
st.header("turn image to audio story")
|
44 |
+
uploaded_file = st.file_uploader("Choose an image ... ", type="jpg")
|
45 |
+
|
46 |
+
if uploaded_file is not None:
|
47 |
+
print(uploaded_file)
|
48 |
+
bytes_data = uploaded_file.getvalue()
|
49 |
+
with open(uploaded_file.name, "wb") as file:
|
50 |
+
file.write(bytes_data)
|
51 |
+
st.image(uploaded_file, caption="Uploaded image", use_column_width=True)
|
52 |
+
text = img_to_text(uploaded_file.name)
|
53 |
+
story = generate_story(text)
|
54 |
+
text_to_speech(story)
|
55 |
+
|
56 |
+
with st.expander("text"):
|
57 |
+
st.write(text)
|
58 |
+
with st.expander("story"):
|
59 |
+
st.write(story)
|
60 |
+
st.audio("audio.flac")
|
61 |
+
|