Spaces:
Running
Running
File size: 2,982 Bytes
c86e71f 3510ec1 | 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 | <!doctype html>
<html>
<head>
<meta charset="utf-8" />
<meta name="viewport" content="width=device-width, initial-scale=1" />
<title>AI Paraphraser - Amna Shahzad</title>
<link rel="stylesheet" href="https://cdn.jsdelivr.net/npm/@stlite/browser@0.85.1/build/stlite.css" />
<link rel="stylesheet" href="https://cdn.jsdelivr.net/npm/@stlite/browser@0.85.1/build/style.css" />
</head>
<body>
<streamlit-app>
import json
import streamlit as st
from pyodide.code import run_js
from pyodide.ffi import to_js
from js import JSON, Object
# ONNX (browser) copy of google/flan-t5-small - T5 text-to-text model
MODEL = "Xenova/flan-t5-small"
st.set_page_config(page_title="AI Paraphraser", page_icon="✨", layout="wide")
# MANDATORY: Name and Student ID in the title
st.title("Amna Shahzad - Student ID: B04-0923-000008")
st.caption("CS4106 Mid-Term | AI Paraphraser (Text-to-Text with T5)")
if "gen" not in st.session_state:
msg = st.empty()
msg.info("Loading T5 model (first run only, please wait)...")
try:
mod = await run_js("import('https://cdn.jsdelivr.net/npm/@huggingface/transformers@3.8.1')")
st.session_state["gen"] = await mod.pipeline("text2text-generation", MODEL)
msg.empty()
except Exception as e:
msg.error(f"Could not load model: {e}")
st.stop()
gen = st.session_state["gen"]
st.success(f"Model loaded: {MODEL}")
text = st.text_area("Enter a sentence to paraphrase:", height=120, max_chars=1000)
creativity = st.slider("Creativity (temperature)", 0.3, 1.5, 0.8)
if st.button("Paraphrase"):
if not text.strip():
st.warning("Please type something first.")
elif len(text.split()) > 150:
st.error("Text too long - please keep it under 150 words.")
else:
try:
with st.spinner("Paraphrasing..."):
prompt = "Paraphrase this sentence in different words: " + text
opts = to_js({"max_new_tokens": 128, "do_sample": True,
"temperature": creativity, "top_p": 0.9},
dict_converter=Object.fromEntries)
raw = await gen(prompt, opts)
output = json.loads(JSON.stringify(raw))[0]["generated_text"]
col1, col2 = st.columns(2)
with col1:
st.subheader("📝 Original")
st.info(text)
st.caption(f"{len(text.split())} words")
with col2:
st.subheader("✨ Paraphrased")
st.success(output)
st.caption(f"{len(output.split())} words")
if output.strip().lower() == text.strip().lower():
st.warning("⚠️ Output is identical to input - try again or raise creativity.")
except Exception as e:
st.error(f"Something went wrong: {e}")
</streamlit-app>
<script type="module" src="https://cdn.jsdelivr.net/npm/@stlite/browser@0.85.1/build/stlite.js"></script>
</body>
</html> |