🌘w🌖

app.py

@@ -76,9 +76,8 @@ def talk(message, history):
     messages = tokenizer.apply_chat_template(
         chat, tokenize=False, add_generation_prompt=True
     )
-    # Initialize an empty string to store the generated text
-    partial_text = ""
-    yield partial_text
+    print("chat template prepared, ", messages)
+    print("tokenizing input ...")
     # Tokenize the messages string
     model_inputs = tokenizer([messages], return_tensors="pt").to(device)
     streamer = TextIteratorStreamer(
@@ -94,11 +93,14 @@ def talk(message, history):
         temperature=0.75,
         num_beams=1,
     )
+    print("initializing thread ...")
     t = Thread(target=model.generate, kwargs=generate_kwargs)
     t.start()
 
+    # Initialize an empty string to store the generated text
+    partial_text = ""
     for new_text in streamer:
-        if new_text is not None :
+        if new_text is not None:
             partial_text += new_text
             yield partial_text
     partial_text += resources