0.13 catch more exceptions

app.py

@@ -24,6 +24,8 @@ models_available = [
     "mistralai/Mistral-7B-Instruct-v0.3",
 ]
 
+device = torch.device("cuda:0" if torch.cuda.is_available() else "cpu")
+
 tokenizer_a, model_a = None, None
 tokenizer_b, model_b = None, None
 torch_dtype = torch.bfloat16
@@ -75,7 +77,6 @@ def load_model_b(model_id):
     try: 
         model_id_b = model_id
         tokenizer_b = AutoTokenizer.from_pretrained(model_id)
-        logging.debug(f"***** model B eos_token: {tokenizer_b.eos_token}")
         model_b = AutoModelForCausalLM.from_pretrained(
             model_id,
             torch_dtype=torch_dtype,
@@ -133,7 +134,7 @@ def generate_both(system_prompt, input_text, chatbot_a, chatbot_b, max_new_token
             top_p=top_p,
             repetition_penalty=repetition_penalty,
         )
-        
+
         generation_kwargs_b = dict(
             input_ids=input_ids_b,
             streamer=text_streamer_b,
@@ -171,6 +172,8 @@ def generate_both(system_prompt, input_text, chatbot_a, chatbot_b, max_new_token
                 yield chatbot_a, chatbot_b
             except StopIteration:
                 finished_a = True
+            except Exception as e:
+                logging.error(f'{SPACER} Error: {e}, Traceback {traceback.format_exc()}')
 
         if not finished_b:
             try:
@@ -183,6 +186,8 @@ def generate_both(system_prompt, input_text, chatbot_a, chatbot_b, max_new_token
                 yield chatbot_a, chatbot_b
             except StopIteration:
                 finished_b = True
+            except Exception as e:
+                logging.error(f'{SPACER} Error: {e}, Traceback {traceback.format_exc()}')
 
     return chatbot_a, chatbot_b