Create app.py

app.py

@@ -0,0 +1,59 @@
+import streamlit as st
+import pandas as pd
+import torch
+from transformers import BertTokenizer, BertForSequenceClassification
+
+# Load pre-trained BERT model and tokenizer
+tokenizer = BertTokenizer.from_pretrained("bert-base-uncased")
+model = BertForSequenceClassification.from_pretrained("bert-base-uncased", num_labels=6)
+
+# Load pre-trained model weights
+model_path = "model.pt"  # Path to pre-trained model weights
+model.load_state_dict(torch.load(model_path, map_location=torch.device('cpu')))
+model.eval()
+
+# Create a DataFrame to store classification results
+results_df = pd.DataFrame(columns=['Text', 'Toxic', 'Obscene', 'Insult', 'Identity_Hate', 'Threat', 'Severe_Toxic'])
+
+def classify_text(text):
+    """
+    Function to classify text using the pre-trained BERT model.
+    """
+    inputs = tokenizer.encode_plus(text, add_special_tokens=True, return_tensors='pt')
+    with torch.no_grad():
+        outputs = model(inputs['input_ids'], attention_mask=inputs['attention_mask'])
+        logits = outputs.logits
+        probabilities = torch.sigmoid(logits)
+    return probabilities[0].tolist()
+
+def main():
+    """
+    Main function to run the Streamlit app.
+    """
+    st.title("Toxicity Classification")
+    text = st.text_area("Enter text to classify", "")
+    if st.button("Classify"):
+        probabilities = classify_text(text)
+        toxic = probabilities[0]
+        obscene = probabilities[1]
+        insult = probabilities[2]
+        identity_hate = probabilities[3]
+        threat = probabilities[4]
+        severe_toxic = probabilities[5]
+        st.write("Toxic: {:.2f}%".format(toxic*100))
+        st.write("Obscene: {:.2f}%".format(obscene*100))
+        st.write("Insult: {:.2f}%".format(insult*100))
+        st.write("Identity Hate: {:.2f}%".format(identity_hate*100))
+        st.write("Threat: {:.2f}%".format(threat*100))
+        st.write("Severe Toxic: {:.2f}%".format(severe_toxic*100))
+        if st.button("Add to Results"):
+            global results_df
+            results_df = results_df.append({'Text': text, 'Toxic': toxic, 'Obscene': obscene,
+                                            'Insult': insult, 'Identity_Hate': identity_hate,
+                                            'Threat': threat, 'Severe_Toxic': severe_toxic}, ignore_index=True)
+            st.success("Text added to Results DataFrame!")
+    if st.button("View Results"):
+        st.write(results_df)
+
+if __name__ == "__main__":
+    main()