Update app.py

app.py

@@ -1,3 +1,4 @@
+# install required packages
 import subprocess
 import sys
 
@@ -8,35 +9,39 @@ install("tensorflow")
 install("numpy")
 install("transformers")
 
-import transformers
-from transformers import DistilBertTokenizer
-from transformers import TFDistilBertForSequenceClassification
-
+# import related packages
 import streamlit as st
 import numpy as np
 import tensorflow as tf
 
+import transformers
+from transformers import DistilBertTokenizer
+from transformers import TFDistilBertForSequenceClassification
+
+# print the header message
 st.header("Welcome to the STEM NLP application!")
+
+# fetch the pre-trained model
 model = TFDistilBertForSequenceClassification.from_pretrained("kaixinwang/NLP")
 
-MODEL_NAME_1 = 'distilbert-base-uncased'
-tokenizer = DistilBertTokenizer.from_pretrained(MODEL_NAME_1)
+# build the tokenizer
+MODEL_NAME = 'distilbert-base-uncased'
+tokenizer = DistilBertTokenizer.from_pretrained(MODEL_NAME)
 
 mapping = {0:"Negative", 1:"Positive"}
+# prompt for the user input
 x = st.text_input("To get started, enter your review/text below and hit ENTER:")
 if x:
-    # st.write("Your review is:", x)
     st.write("Determining the sentiment...")
+    # utterance tokenization
     encoding = tokenizer([x], truncation=True, padding=True)
     encoded = tf.data.Dataset.from_tensor_slices((dict(encoding), np.ones(1)))
+    # make the prediction
     preds = model.predict(encoded.batch(1)).logits  
     prob = tf.nn.softmax(preds, axis=1).numpy()  
     prob_max = np.argmax(prob, axis=1)
+    # display the output
     st.write("Your review is:", x)
     content = "Sentiment: %s, prediction score: %.4f" %(mapping[prob_max[0]], prob[0][prob_max][0])
     st.write(content)
-    # st.write("Sentiment:", mapping[prob_max[0]], "Prediction Score:", prob[0][prob_max][0])
-
-
-# x = st.slider('Select a value')
-# st.write(x, 'squared is', x * x)
+    # st.write("Sentiment:", mapping[prob_max[0]], "Prediction Score:", prob[0][prob_max][0])