File size: 2,100 Bytes
c395f90 60e3188 c395f90 cb5aeb3 682ebcc cb5aeb3 60e3188 2b2e0e0 60e3188 2b2e0e0 cb5aeb3 2b2e0e0 682ebcc cb5aeb3 682ebcc cb5aeb3 682ebcc 60e3188 cb5aeb3 682ebcc 60e3188 682ebcc | 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 | import os
# ✅ Fix: Use writable directories for cache and metrics
os.environ["TRANSFORMERS_CACHE"] = "/tmp/hf"
os.environ["HF_HOME"] = "/tmp/hf"
os.environ["STREAMLIT_HOME"] = "/tmp/.streamlit"
os.makedirs("/tmp/hf", exist_ok=True)
os.makedirs("/tmp/.streamlit", exist_ok=True)
import streamlit as st
from transformers import pipeline
# 🔐 Load Hugging Face token from Space secrets
HF_TOKEN = os.getenv("HF_TOKEN")
# 🔁 Load the private model using the token
classifier = pipeline(
"text-classification",
model="azratuni/isl-classifier",
token=HF_TOKEN,
return_all_scores=True # ✅ Return all class scores
)
# 🧠 Streamlit UI
st.set_page_config(page_title="Islamophobia Classifier", page_icon="🛡️")
st.title("🛡️ Islamophobia Detection")
st.write("Enter a sentence below to check if it's Islamophobic or not.")
text_input = st.text_area("Input text", placeholder="e.g. Muslims are terrorists.")
if st.button("Classify"):
if text_input.strip():
with st.spinner("Classifying..."):
scores = classifier(text_input)[0]
scores = sorted(scores, key=lambda x: x["score"], reverse=True)
top_label = scores[0]["label"]
top_score = scores[0]["score"]
second_score = scores[1]["score"]
margin = top_score - second_score
label_map = {
"LABEL_0": "Not Islamophobic",
"LABEL_1": "Islamophobic"
}
label = label_map.get(top_label, top_label)
st.success(f"**Prediction:** {label}")
st.write(f"**Confidence:** {top_score:.2%}")
# ⚠️ Uncertainty warning if margin is low
if margin < 0.15:
st.warning("⚠️ The model is uncertain. Both classes received similar confidence scores.")
# 🧾 Optional: show both scores
st.markdown("**Score breakdown:**")
for s in scores:
st.write(f"- {label_map.get(s['label'], s['label'])}: {s['score']:.2%}")
else:
st.warning("Please enter some text.")
|