Upload app.py

app.py

@@ -0,0 +1,35 @@
+import streamlit as st
+from transformers import AutoModelForSequenceClassification, AutoTokenizer
+import torch
+
+# Specify the local paths where the model and tokenizer are saved
+model_dir = "path_to_save_model"
+tokenizer_dir = "path_to_save_tokenizer"
+
+# Load the model and tokenizer from the local directories
+model = AutoModelForSequenceClassification.from_pretrained(model_dir)
+tokenizer = AutoTokenizer.from_pretrained(tokenizer_dir)
+
+# Function to predict the sentiment
+def predict_sentiment(text):
+    inputs = tokenizer(text, return_tensors="pt")
+    with torch.no_grad():
+        outputs = model(**inputs)
+    probs = torch.nn.functional.softmax(outputs.logits, dim=-1)
+    return torch.argmax(probs, dim=-1).item(), probs
+
+# Streamlit interface
+st.title("KIWI Classifier")
+st.write("Enter a question or statement to classify:")
+
+user_input = st.text_area("Your input", "")
+if st.button("Classify"):
+    if user_input:
+        label, probabilities = predict_sentiment(user_input)
+        st.write(f"Prediction: {label}")
+        st.write(f"Probabilities: {probabilities.tolist()}")
+    else:
+        st.write("Please enter some text to classify.")
+
+# Additional instructions or information
+st.write("This application uses a fine-tuned BERT model to classify questions and statements.")