Spaces:
Sleeping
Sleeping
Commit
•
33903b2
1
Parent(s):
a347a3b
Upload app.py
Browse files
app.py
ADDED
@@ -0,0 +1,83 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
from transformers import pipeline
|
2 |
+
from dotenv import find_dotenv, load_dotenv
|
3 |
+
from langchain import PromptTemplate, LLMChain, HuggingFaceHub
|
4 |
+
import streamlit as st
|
5 |
+
import requests
|
6 |
+
import os
|
7 |
+
|
8 |
+
|
9 |
+
load_dotenv(find_dotenv())
|
10 |
+
HUGGINGFACE_API = os.getenv("HUGGINGFACE_API")
|
11 |
+
|
12 |
+
|
13 |
+
def image2text(url):
|
14 |
+
image_to_text = pipeline('image-to-text', model='Salesforce/blip-image-captioning-large')
|
15 |
+
text = image_to_text(url)[0]['generated_text']
|
16 |
+
print(text)
|
17 |
+
return text
|
18 |
+
|
19 |
+
|
20 |
+
|
21 |
+
def generate_story(scenario, length):
|
22 |
+
template = """
|
23 |
+
You are story teller, generate a short story in {length} words\n
|
24 |
+
CONTEXT:{scenario}\n
|
25 |
+
STORY:
|
26 |
+
"""
|
27 |
+
|
28 |
+
prompt = PromptTemplate(template=template, input_variables=["scenario","length"])
|
29 |
+
llm = LLMChain(llm=HuggingFaceHub(huggingfacehub_api_token=HUGGINGFACE_API, repo_id="mistralai/Mixtral-8x7B-Instruct-v0.1"), prompt=prompt, verbose=True)
|
30 |
+
story = llm.predict(scenario=scenario, length=length)
|
31 |
+
print(story)
|
32 |
+
return story
|
33 |
+
|
34 |
+
|
35 |
+
|
36 |
+
# def text2speech(message):
|
37 |
+
# API_URL = "https://api-inference.huggingface.co/models/microsoft/speecht5_tts"
|
38 |
+
# headers = {"Authorization": f"Bearer {HUGGINGFACE_API}"}
|
39 |
+
# payloads = {
|
40 |
+
# "inputs": message
|
41 |
+
# }
|
42 |
+
# response = requests.post(API_URL,headers=headers,json=payloads)
|
43 |
+
# with open('audio.wav', 'wb') as file:
|
44 |
+
# file.write(response.content)
|
45 |
+
|
46 |
+
|
47 |
+
|
48 |
+
def main():
|
49 |
+
st.set_page_config(page_title="Image Storyteller")
|
50 |
+
|
51 |
+
st.header("Image to Audio")
|
52 |
+
uploaded_file = st.file_uploader("Choose an Image", type="jpg")
|
53 |
+
|
54 |
+
length = st.number_input("Length")
|
55 |
+
scenario = ""
|
56 |
+
successful_processing = False
|
57 |
+
|
58 |
+
if uploaded_file is not None:
|
59 |
+
print(uploaded_file)
|
60 |
+
bytes_data = uploaded_file.getvalue()
|
61 |
+
with open(uploaded_file.name, "wb") as file:
|
62 |
+
file.write(bytes_data)
|
63 |
+
|
64 |
+
st.image(uploaded_file.name, caption="Uploaded Image", use_column_width=True)
|
65 |
+
|
66 |
+
try:
|
67 |
+
scenario = image2text(uploaded_file.name)
|
68 |
+
successful_processing = True
|
69 |
+
except Exception as e:
|
70 |
+
st.error(f"Error processing the image: {e}")
|
71 |
+
|
72 |
+
if successful_processing:
|
73 |
+
story = generate_story(scenario, length)
|
74 |
+
# text2speech(story)
|
75 |
+
|
76 |
+
with st.expander("scenario"):
|
77 |
+
st.write(scenario)
|
78 |
+
with st.expander("generated story"):
|
79 |
+
st.write(story)
|
80 |
+
# st.audio('audio.wav')
|
81 |
+
|
82 |
+
if __name__ == '__main__':
|
83 |
+
main()
|