Update app.py

app.py

@@ -3,10 +3,11 @@ from gradio import mix
 import tensorflow as tf
 import numpy as np
 import torch
-from transformers import AutoTokenizer, AutoModelForSeq2SeqLM
+from keras.preprocessing.sequence import pad_sequences
+import pickle
+
 from huggingface_hub import from_pretrained_keras
 
-tokenizer = AutoTokenizer.from_pretrained("bert-base-cased")
 model = from_pretrained_keras("keras-io/text-generation-miniature-gpt")
 
 def tokenize_data(text):
@@ -21,14 +22,14 @@ def tokenize_data(text):
     return inputs
 
 def generate_answers(text):
-    inputs = tokenize_data(text)
-    x = tf.dtypes.cast(inputs['input_ids'], tf.int64)
-    predictions, _ = model.predict(x)
-    #predictions, _ = model.predict(inputs['input_ids'])
-    results = np.argmax(predictions, axis=1)
-    answer = tokenizer.decode(results[0].flatten(), skip_special_tokens=True)
-    return answer
-
+    sequence_test = tokenizer.texts_to_sequences([text])
+    padded_test = pad_sequences(sequence_test, maxlen= 80, padding='post')
+    predictions,_ = model.predict(padded_test)
+    results = np.argmax(predictions, axis=1)[0]
+    answer = tokenizer.sequences_to_texts([results] )
+    answertoString = ' '.join([str(elem) for elem in answer])
+    return answertoString
+    
 examples = [["Once upon a time, "], ["In a galaxy far far away"]]
 title = "Text Generation with Miniature GPT"
 description = "Gradio Demo for a miniature with GPT. To use it, simply add your text, or click one of the examples to load them. Read more at the links below."