File size: 816 Bytes
9688990 | 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 | #load model
from tensorflow.keras.models import load_model
from tensorflow.keras.preprocessing.sequence import pad_sequences
import pickle
model = load_model("model.h5")
#load tokenizer
with open("tokenizer.pkl","rb") as handle:
tokenizer = pickle.load(handle)
#make predictions
# Make predictions
while True:
text = input("write a review, press e to exit: ")
if text == 'e':
break
TokenText = tokenizer.texts_to_sequences([text])
PadText = pad_sequences(TokenText, maxlen=100)
Pred = model.predict(PadText)
Pred_float = Pred[0][0] # Extract the single float value
Pred_float *= 1.3
binary_pred = (Pred_float > 0.5).astype(int)
if binary_pred == 0:
print("bad review")
else:
print("good review")
print(Pred_float) |