Create app.py

app.py

@@ -0,0 +1,38 @@
+import tensorflow as tf
+from tensorflow.keras.models import load_model
+import json
+import keras_nlp
+import gradio as gr
+
+fnet_classifier = load_model("Sentiments classifier.keras")
+
+with open("vocab.json", "r") as f:
+    vocab = json.load(f)
+
+seq_max_length = 512
+tokenizer = keras_nlp.tokenizers.WordPieceTokenizer(
+    vocabulary=vocab,
+    lowercase=False,
+    sequence_length=seq_max_length,
+)
+
+def make_prediction(sentence):
+    tokens = tokenizer(review_example)
+    tokens = tf.expand_dims(tokens, 0)
+    prediction = fnet_classifier.predict(tokens, verbose=0)
+
+    if prediction[0][0] > 0.5:
+        result = "The review is POSITIVE"
+    else:
+        result = "The review is NEGATIVE"
+    return result
+
+gradio_app = gr.Interface(
+    make_prediction,
+    inputs=gr.Textbox(label="Your review"),
+    outputs=[gr.Textbox(label="Sentiment"),
+    title="Positive Review or Negtaive Review",
+)
+
+if __name__ == "__main__":
+    gradio_app.launch()