Spaces:
Sleeping
Sleeping
File size: 2,536 Bytes
69bf39e 5e9dd30 69bf39e 249d7ba 69bf39e 62e64fb 5e9dd30 62e64fb 5e9dd30 62e64fb 5e9dd30 62e64fb 5e9dd30 62e64fb 5e9dd30 249d7ba 5e9dd30 249d7ba 985d58e 249d7ba 5e9dd30 249d7ba 5e9dd30 249d7ba d9e1771 249d7ba 5e9dd30 249d7ba 7bdff6e 249d7ba 7ff270d 249d7ba 985d58e 249d7ba 6feb14e 62e64fb |
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 72 73 74 75 76 77 78 79 80 |
import streamlit as st
from PyPDF2 import PdfReader
from transformers import AutoTokenizer, AutoModelForCausalLM
import torch
from io import BytesIO
# Initialize the tokenizer and model
tokenizer = AutoTokenizer.from_pretrained("himmeow/vi-gemma-2b-RAG")
model = AutoModelForCausalLM.from_pretrained(
"himmeow/vi-gemma-2b-RAG",
device_map="auto",
torch_dtype=torch.bfloat16
)
# Use GPU if available
if torch.cuda.is_available():
model.to("cuda")
# Streamlit app layout
st.set_page_config(page_title="π PDF Query App", page_icon=":book:", layout="wide")
st.title("π PDF Query App")
st.sidebar.title("Upload File and Query")
# Sidebar: File Upload
uploaded_file = st.sidebar.file_uploader("Upload your PDF file", type="pdf")
# Sidebar: Query Input
query = st.sidebar.text_input("Enter your query:")
# Sidebar: Submit Button
if st.sidebar.button("Submit"):
if uploaded_file and query:
# Read the PDF file
pdf_text = ""
with BytesIO(uploaded_file.read()) as file:
reader = PdfReader(file)
for page_num in range(len(reader.pages)):
page = reader.pages[page_num]
text = page.extract_text()
pdf_text += text + "\n"
# Define the prompt format for the model
prompt = """
{}
Please answer the question: {}
{}
"""
# Format the input text
input_text = prompt.format(pdf_text, query, " ")
# Encode the input text into input ids
input_ids = tokenizer(input_text, return_tensors="pt")
# Use GPU for input ids if available
if torch.cuda.is_available():
input_ids = input_ids.to("cuda")
# Generate text using the model
outputs = model.generate(
**input_ids,
max_new_tokens=500, # Limit the number of tokens generated
no_repeat_ngram_size=5, # Prevent repetition of 5-gram phrases
)
# Decode and display the results
response = tokenizer.decode(outputs[0], skip_special_tokens=True)
# Remove unwanted text fields from the response
clean_response = response.replace("### Instruction and Input:", "").replace("### Response:", "").strip()
st.write(clean_response)
else:
st.sidebar.error("Please upload a PDF file and enter a query.")
# Footer with LinkedIn link
st.sidebar.write("---")
st.sidebar.write("Created by: [Engr. Hamesh Raj](https://www.linkedin.com/in/datascientisthameshraj/)") |