Update app.py

app.py

@@ -1,11 +1,21 @@
+import os
 import gradio as gr
 import torch
 from sentence_transformers import SentenceTransformer, util
 
 # -----------------------------
-# Load model (FAST)
+# Hard CPU enforcement (HF-safe)
 # -----------------------------
-model = SentenceTransformer("all-MiniLM-L6-v2")
+os.environ["CUDA_VISIBLE_DEVICES"] = ""
+torch.set_num_threads(1)
+
+# -----------------------------
+# Load model (FAST & SAFE)
+# -----------------------------
+model = SentenceTransformer(
+    "sentence-transformers/all-MiniLM-L6-v2",
+    device="cpu"
+)
 
 # -----------------------------
 # Intent anchors
@@ -14,7 +24,6 @@ INTENT_LABELS = {
     "chat": [
         "casual conversation",
         "small talk",
-        "talk to me",
         "just chatting"
     ],
     "search": [
@@ -29,33 +38,29 @@ INTENT_LABELS = {
     ],
     "code": [
         "write code",
-        "debug program",
+        "debug a program",
         "programming help"
     ],
     "research": [
         "deep research",
-        "academic analysis",
-        "investigate topic"
+        "academic analysis"
     ],
     "study": [
         "help me learn",
-        "teach me something",
         "study help"
     ],
     "project": [
         "build a project",
-        "create an application",
-        "long term development"
+        "create an application"
     ],
     "action": [
         "set a reminder",
-        "schedule something",
         "perform an action"
     ]
 }
 
 # -----------------------------
-# Precompute anchor embeddings
+# Precompute embeddings (ONCE)
 # -----------------------------
 anchor_embeddings = {
     intent: model.encode(phrases, convert_to_tensor=True)
@@ -65,16 +70,16 @@ anchor_embeddings = {
 # -----------------------------
 # Classifier
 # -----------------------------
-def classify(text: str):
+def classify(text):
     if not text.strip():
-        return {"error": "Empty input"}
+        return {}
 
     query_emb = model.encode(text, convert_to_tensor=True)
 
-    scores = {}
-    for intent, emb in anchor_embeddings.items():
-        score = util.cos_sim(query_emb, emb).max().item()
-        scores[intent] = score
+    scores = {
+        intent: util.cos_sim(query_emb, emb).max().item()
+        for intent, emb in anchor_embeddings.items()
+    }
 
     top_intent = max(scores, key=scores.get)
 
@@ -92,7 +97,16 @@ demo = gr.Interface(
     inputs=gr.Textbox(label="User Input"),
     outputs=gr.JSON(),
     title="⚡ Ultra-Fast Intent Router",
-    description="Sub-second semantic routing using MiniLM (no training)"
+    description="MiniLM semantic router (no training, sub-second)"
 )
 
-app = demo
+# -----------------------------
+# REQUIRED FOR HF SPACES
+# -----------------------------
+if __name__ == "__main__":
+    demo.launch(
+        server_name="0.0.0.0",
+        server_port=7860,
+        share=True,
+        show_error=True
+    )