DrishtiSharma's picture
Update app.py
8e4cbbb
import gradio as gr
import tensorflow as tf
from tensorflow.keras.preprocessing.sequence import pad_sequences
import pickle
from huggingface_hub import from_pretrained_keras
import os
os.environ['TF_CPP_MIN_LOG_LEVEL'] = '2'
model = from_pretrained_keras("keras-io/bidirectional-lstm-imdb")
with open('tokenizer.pickle', 'rb') as file:
tokenizer = pickle.load(file)
def decide(text):
tokenized_text = tokenizer.texts_to_sequences([text])
padded_tokens = pad_sequences(tokenized_text, maxlen= 200)
result = model.predict(padded_tokens)[0][0]
if result >= 0.6 :
return "Positive review"
elif result <= 0.4:
return "Negative review"
else:
return "Neutral review"
example_sentence_1 = "I hate the movie, they made no effort in making the movie. Waste of time!"
example_sentence_2 = "Awesome movie! Loved the way in which the hero acted."
examples = [[example_sentence_1], [example_sentence_2]]
description = "Write out a movie review to know the underlying sentiment."
article = "<div style='text-align: center;'><a href='https://huggingface.co/DrishtiSharma' target='_blank'>Space by Drishti Sharma</a><br><a href='https://keras.io/examples/nlp/bidirectional_lstm_imdb/' target='_blank'>Keras example by François Chollet</a></div>"
gr.Interface(decide, inputs= gr.inputs.Textbox( lines=1, placeholder=None, default="", label=None), outputs='text', examples=examples,
title="Sentiment analysis of movie reviews", theme = "grass", description=description, allow_flagging="auto",
flagging_dir='flagging records', article = article).launch(enable_queue = True,
inline=False, share = True)