Spaces:
Runtime error
Runtime error
Update app.py
Browse files
app.py
CHANGED
|
@@ -1,43 +1,30 @@
|
|
| 1 |
import streamlit as st
|
| 2 |
from transformers import pipeline
|
| 3 |
from PIL import Image
|
| 4 |
-
import requests
|
| 5 |
-
|
| 6 |
-
# Hardcoded OpenAI API Key
|
| 7 |
-
OPENAI_API_KEY = "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"
|
| 8 |
|
| 9 |
# Load OCR model for extracting text from images
|
| 10 |
@st.cache_resource
|
| 11 |
def load_ocr_model():
|
| 12 |
return pipeline("image-to-text", model="Salesforce/blip-image-captioning-base")
|
| 13 |
|
| 14 |
-
#
|
| 15 |
-
|
| 16 |
-
|
| 17 |
-
|
| 18 |
-
|
| 19 |
-
|
| 20 |
-
|
| 21 |
-
|
| 22 |
-
|
| 23 |
-
|
| 24 |
-
"max_tokens": 1000,
|
| 25 |
-
"temperature": 0.7,
|
| 26 |
-
}
|
| 27 |
-
response = requests.post(url, headers=headers, json=data)
|
| 28 |
-
if response.status_code == 200:
|
| 29 |
-
return response.json()["choices"][0]["message"]["content"].strip()
|
| 30 |
-
else:
|
| 31 |
-
st.error(f"Error from OpenAI: {response.status_code} - {response.text}")
|
| 32 |
-
return None
|
| 33 |
|
| 34 |
# Streamlit App
|
| 35 |
def main():
|
| 36 |
-
st.title("Image-to-Text with
|
| 37 |
st.markdown(
|
| 38 |
"""
|
| 39 |
-
**Upload an image**, extract text using
|
| 40 |
-
and get explanations or
|
| 41 |
"""
|
| 42 |
)
|
| 43 |
|
|
@@ -59,9 +46,9 @@ def main():
|
|
| 59 |
st.write("### Extracted Text:")
|
| 60 |
st.write(f"`{extracted_text}`") # Display the extracted text in a readable format
|
| 61 |
|
| 62 |
-
#
|
| 63 |
-
st.write("###
|
| 64 |
-
explanation =
|
| 65 |
if explanation:
|
| 66 |
st.write(explanation)
|
| 67 |
else:
|
|
|
|
| 1 |
import streamlit as st
|
| 2 |
from transformers import pipeline
|
| 3 |
from PIL import Image
|
|
|
|
|
|
|
|
|
|
|
|
|
| 4 |
|
| 5 |
# Load OCR model for extracting text from images
|
| 6 |
@st.cache_resource
|
| 7 |
def load_ocr_model():
|
| 8 |
return pipeline("image-to-text", model="Salesforce/blip-image-captioning-base")
|
| 9 |
|
| 10 |
+
# Load text-generation model
|
| 11 |
+
@st.cache_resource
|
| 12 |
+
def load_text_model():
|
| 13 |
+
return pipeline("text-generation", model="EleutherAI/gpt-neo-1.3B")
|
| 14 |
+
|
| 15 |
+
# Function to process text with a language model
|
| 16 |
+
def process_with_llm(prompt):
|
| 17 |
+
llm_model = load_text_model()
|
| 18 |
+
response = llm_model(prompt, max_length=500, do_sample=True, temperature=0.7)
|
| 19 |
+
return response[0]["generated_text"]
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 20 |
|
| 21 |
# Streamlit App
|
| 22 |
def main():
|
| 23 |
+
st.title("Image-to-Text with Open-Source Language Models")
|
| 24 |
st.markdown(
|
| 25 |
"""
|
| 26 |
+
**Upload an image**, extract text using an open-source OCR model,
|
| 27 |
+
and get explanations or text completions using a GPT-style open-source model.
|
| 28 |
"""
|
| 29 |
)
|
| 30 |
|
|
|
|
| 46 |
st.write("### Extracted Text:")
|
| 47 |
st.write(f"`{extracted_text}`") # Display the extracted text in a readable format
|
| 48 |
|
| 49 |
+
# Process extracted text with LLM
|
| 50 |
+
st.write("### Explanation/Completion:")
|
| 51 |
+
explanation = process_with_llm(extracted_text)
|
| 52 |
if explanation:
|
| 53 |
st.write(explanation)
|
| 54 |
else:
|