fix: no need to remove last dialogue

agents/bart_agent.py

@@ -50,7 +50,7 @@ class BARTResponseAgent(object):
         persona = [tokenize(line.strip()) for line in conversation['persona B']]
         # partner = [tokenize(line.strip()) for line in conversation['persona A']]
         partner = [] # Baseline not trained with the partner personaj
-        history = [tokenize(line['text'].strip()) for line in conversation['dialogue'][:-1]]
+        history = [tokenize(line['text'].strip()) for line in conversation['dialogue']]
         return persona, partner, history
    
     def prepare_tensors(self, conversation):