Spaces:
Running
Running
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(sentence) | |
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() |