Create app.py

app.py

@@ -0,0 +1,63 @@
+import gradio as gr
+import joblib
+import xgboost as xgb
+from transformers import AutoModelForSequenceClassification, AutoTokenizer
+import torch
+import numpy as np
+
+# Load BERT model & tokenizer
+tokenizer = AutoTokenizer.from_pretrained("distilbert-base-cased")
+model = AutoModelForSequenceClassification.from_pretrained('distilbert-base-cased', num_labels=5,output_hidden_states=True,trust_remote_code=True)
+from safetensors.torch import load_file
+state_dict = load_file("model (4).safetensors")
+
+# Load the state into the model
+model.load_state_dict(state_dict,strict=False)
+model.eval()
+
+# Load PCA and Scaler
+pca = joblib.load("pca.pkl")
+scaler = joblib.load("scaler.pkl")
+kmean=joblib.load("kmeans_model.pkl")
+
+# Load XGBoost model
+xgb_model = xgb.XGBClassifier()
+xgb_model.load_model("xgb_model.json")
+
+category_mappings = {
+    "kmeans_labels": pd.Index([0, 1, 2, 3, 4, 5, 6, 7, 8, 9], dtype='int32')
+}
+
+def infer(component,title,description):
+    # BERT embedding
+    combined_text = f"{component} [SEP] {title} [SEP] {description}"
+    inputs = tokenizer(combined_text, return_tensors="pt", truncation=True,max_length=512)
+    with torch.no_grad():
+        outputs = model(**inputs)
+    cls_embedding = outputs.hidden_states[-1][:, 0, :].numpy()
+    test_df = pd.DataFrame(cls_embedding)
+
+    # Preprocessing
+    test_pca = scaler.transform(test_df)
+    test_pca = pca.transform(test_pca)
+    test_df = pd.DataFrame(test_pca, columns=[f"PCA{i+1}" for i in range(n)], index=test_df.index)
+    kmeans_labels_test = kmeans.predict(test_df)
+    test_df["kmeans_labels"]=kmeans_labels_test
+    test_df["kmeans_labels"] = pd.Categorical(test_df["kmeans_labels"], categories=category_mappings["kmeans_labels"])
+    
+
+    # Predict
+    prediction = xgb_model.predict(test_df,iteration_range=(0, 130))
+    return f"Predicted Priority: {int(prediction[0])}"
+
+# Gradio interface
+iface = gr.Interface(
+    fn=infer,
+    inputs=[
+        gr.Textbox(label="Component"),
+        gr.Textbox(label="Title"),
+        gr.Textbox(label="Description")
+    ],
+    outputs="text"
+)
+iface.launch()