BSrikanth commited on
Commit
78f9150
·
1 Parent(s): f4439c9

Delete app.py

Browse files
Files changed (1) hide show
  1. app.py +0 -70
app.py DELETED
@@ -1,70 +0,0 @@
1
- from dotenv import find_dotenv, load_dotenv
2
- from transformers import pipeline
3
- from langchain import PromptTemplate, LLMChain, OpenAI
4
- import requests
5
- import os
6
- import streamlit as st
7
-
8
- load_dotenv(find_dotenv())
9
- HUGGINGPHASE_API_TOKEN = os.getenv("HUGGINGPHASE_API_TOKEN")
10
-
11
- # Image2Text
12
-
13
- def img2text(url):
14
- image_to_text = pipeline("image-to-text", model="Salesforce/blip-image-captioning-large")
15
- text = img2text(url)[0]["generated_text"]
16
- print(text)
17
- return text
18
-
19
- # LLM
20
-
21
- def generate_story(scenario):
22
- template = """
23
- you are a story teller
24
- you can generate a short story based on simple narrative, the story should be more than 20 words;
25
- CONTEXT:{scenario},
26
- STORY:
27
- """
28
- prompt = PromptTemplate(template=template, input_variable=["scenario"])
29
- story_llm = LLMChain(llm=OpenAI(model_name="gpt-3.5-turbo", temperature=1), prompt=prompt, verbose=True)
30
- story = story_llm.predict(scenario=scenario)
31
- print(story)
32
- return story
33
-
34
- # Text to Speech
35
-
36
- def text2speech(message):
37
- API_URL = "https://api-inference.huggingface.co/models/espnet/kan-bayashi_ljspeech_vits"
38
- headers = {"Authorization": f"Bearer {HUGGINGPHASE_API_TOKEN}"}
39
- payloads = {
40
- "inputs": message
41
- }
42
-
43
- response = requests.post(API_URL, headers=headers, json=payloads)
44
-
45
- with open('audio.flac', 'wb') as file:
46
- file.write(response.content)
47
-
48
- def main():
49
- st.set_page_config(page_title="img 2 Audio story", page_icon='🤖')
50
- st.header("Turn img into an audio story")
51
- uploaded_file = st.file_uploader("Choose an image...", type="jpg")
52
-
53
- if uploaded_file is not None:
54
- print(uploaded_file)
55
- with open(uploaded_file.name, "wb") as file:
56
- file.write(uploaded_file.getvalue())
57
- st.image(uploaded_file, caption="Uploaded Image.", use_column_width=True)
58
- scenario = img2text(uploaded_file.name)
59
- story = generate_story(scenario)
60
- text2speech(story)
61
-
62
- with st.expander("Scenario"):
63
- st.write(scenario)
64
- with st.expander("Story"):
65
- st.write(story)
66
-
67
- st.audio("audio.flac")
68
-
69
- if __name__ == '__main__':
70
- main()