add timing logs & reduce tokens

app.py

@@ -65,13 +65,28 @@ with col2:
         if not title.strip():
             st.warning("Please enter a question first.")
         else:
+            # 1) mark the start
+            t0 = time.time()
+            st.write(f"⏱ Retrieval start: {t0:.1f}s")
+
+            # retrieval
             q_emb = _encoder.encode(title, convert_to_tensor=True)
             sims = torch.nn.functional.cosine_similarity(q_emb.unsqueeze(0), passage_embs)
             topk = torch.topk(sims, k=min(3, len(passages))).indices.tolist()
             context = "\n\n".join(passages[i] for i in topk)
-            prompt = f"Use these notes to answer the question:\n\n{context}\n\nQ: {title}\nA:"
-            with st.spinner("Retrieving & generating…"):
-                out = deepseek_gen(prompt, max_new_tokens=100, do_sample=False)
+
+            t1 = time.time()
+            st.write(f"⏱ Retrieval done in {t1-t0:.1f}s; generation starting…")
+
+            # 2) generation (reduce tokens for now)
+            out = deepseek_gen(
+                f"Use these notes to answer:\n\n{context}\n\nQ: {title}\nA:",
+                max_new_tokens=20,
+                do_sample=False
+            )
+
+            t2 = time.time()
+            st.write(f"⏱ Generation took {t2-t1:.1f}s (total {t2-t0:.1f}s)")
             st.write(out[0]["generated_text"])
 
 # Embedding Q&A in col3