Update app.py

app.py

@@ -71,6 +71,8 @@ class StopOnTokens(StoppingCriteria):
 def predict(message, history):
     text = vector_search(message)
     message = message + "\n\n### Source ###\n" + text
+
+    print(message)
     history_transformer_format = history + [[message, ""]]
     stop = StopOnTokens()
 
@@ -85,11 +87,9 @@ def predict(message, history):
         model_inputs,
         streamer=streamer,
         max_new_tokens=1024,
-        do_sample=True,
+        do_sample=False,
         top_p=0.95,
-        top_k=1000,
-        temperature=1.0,
-        num_beams=1,
+        temperature=0.4,
         stopping_criteria=StoppingCriteriaList([stop])
         )
     t = Thread(target=model.generate, kwargs=generate_kwargs)