Fix of repeating inputs

app.py

@@ -6,6 +6,7 @@ from transformers import AutoModelForSeq2SeqLM, AutoModelForCausalLM
 from transformers import AutoTokenizer
 import numpy as np
 import time
+import string
 
 # JS
 import nltk
@@ -37,6 +38,15 @@ def get_models(llama=False):
 
 model, tokenizer = get_models()
 
+def remove_punctuation(word):
+    # Create a translation table that maps all punctuation characters to None
+    translator = str.maketrans('', '', string.punctuation)
+    
+    # Use the translate method to remove punctuation from the word
+    word_without_punctuation = word.translate(translator)
+    
+    return word_without_punctuation
+
 def return_top_k(sentence, k=10):
   
   if sentence[-1] != ".":
@@ -64,9 +74,10 @@ def return_top_k(sentence, k=10):
     #all word predictions
     predictions = [tokenizer.decode(tokens, skip_special_tokens=True) for tokens in output_sequences['sequences']]
     probabilities = [round(float(prob), 2) for prob in decoded_probabilities]
-    
+
+    stripped_sent = [remove_punctuation(word) for word in sentence.split()]
     for pred in predictions:
-      if (len(pred) < 2) | (pred in sentence.split()):
+      if (len(pred) < 2) | (pred in stripped_sent):
         predictions.pop(predictions.index(pred))  
      
   return predictions[:10]