3

main.py

@@ -7,59 +7,24 @@ from tensorflow.keras import callbacks as kc
 from tensorflow.keras.preprocessing.text import Tokenizer as kT
 from tensorflow.keras.utils import pad_sequences as kps
 
-import re
 import numpy as np
 import pandas as pd
 
-import nltk
-from nltk.corpus import stopwords
-from nltk.tokenize import word_tokenize
-from nltk.stem import WordNetLemmatizer
-
-from tqdm import tqdm
-tqdm.pandas()
-
-
 print("I")
-t = pd.read_csv("yelp_review_polarity_csv/train.csv",
+t = pd.read_csv("prepped_train.csv",
                 header = None,
-                names = ['c', 'r'])
+                names = ['i', 'c', 'r'])
 print("R")
 
-y = t['c'] - 1
+y = t['c']
-r = t['r']
+r = t['r'].astype(str)
-
-#nltk.download("stopwords")
-#nltk.download("punkt_tab")
-#nltk.download("wordnet")
-
-def fr(r):
-    r = r.lower()
-
-    r = " ".join(tuple(re.findall(r'\w+', r)))
-
-    for i in ['\n', '\r', ',', '.', '-', ';', ':', '\'', '"']:
-        r = r.replace(i, "")
-
-    sw = set(stopwords.words("english"))
-    
-    l = WordNetLemmatizer()
-    return " ".join([l.lemmatize(i.strip(), pos = 'v') for i in word_tokenize(r) if i.strip() not in sw])
-
-r = r.progress_apply(fr)
-#print(r)
-
-print("A")
 
 tk = kT(num_words = 6000)
-
 tk.fit_on_texts(r)
 
 print("F")
-#print(tk.word_index)
 
 s = tk.texts_to_sequences(r)
-#print(s)
 
 print("T")
 
@@ -67,6 +32,7 @@ ts = kps(s, maxlen = 100)
 
 print("P")
 
+'''
 m = km.Sequential([
     kl.Input(shape = (None, ), dtype = 'int32'),
     kl.Embedding(6000, 96),
@@ -82,17 +48,19 @@ m = km.Sequential([
 m.compile(optimizer = ko.Lion(learning_rate = 0.0005),
           loss = 'binary_crossentropy',
           metrics = ['accuracy'])
+'''
 
-#m.summary()
+from model import m
 
-ckpt = kc.ModelCheckpoint('model1.keras',
+ckpt = kc.ModelCheckpoint('model2.keras',
                           monitor = 'val_accuracy',
                           save_best_only = True,
                           verbose = 1)
 
+m.load_weights("model1.keras")
 history = m.fit(ts,
                 y,
-                epochs = 3,
+                epochs = 15,
-                batch_size = 256,
+                batch_size = 1024,
                 validation_split = 0.1,
                 callbacks = [ckpt])

model.py

@@ -0,0 +1,21 @@
+from tensorflow.keras import layers as kl
+from tensorflow.keras import models as km
+from tensorflow.keras import losses as ks
+from tensorflow.keras import optimizers as ko
+from tensorflow.keras import callbacks as kc
+
+m = km.Sequential([
+    kl.Input(shape = (None, ), dtype = 'int32'),
+    kl.Embedding(6000, 96),
+    kl.Dropout(0.2),
+    kl.Conv1D(128, 5, activation = 'relu'),
+    kl.LSTM(128, return_sequences = True),
+    kl.LSTM(64),
+    kl.Dense(64),
+    kl.Dropout(0.5),
+    kl.Dense(1, activation = 'sigmoid')
+])
+
+m.compile(optimizer = ko.Lion(learning_rate = 0.0005),
+          loss = 'binary_crossentropy',
+          metrics = ['accuracy'])

plot.py

@@ -0,0 +1,11 @@
+from matplotlib import pyplot as plt
+
+X = [i for i in range(1, 21)]
+tY = [0.8380, 0.9170, 0.9267, 0.9333, 0.9390, 0.9373, 0.9457, 0.9541, 0.9609, 0.9661, 0.9703, 0.9732, 0.9757, 0.9775, 0.9786, 0.9799, 0.9810, 0.9820, 0.9822, 0.9829]
+vY = [0.9242, 0.9346, 0.9392, 0.9414, 0.9436, 0.9452, 0.9453, 0.9457, 0.9441, 0.9432, 0.9423, 0.9429, 0.9415, 0.9420, 0.9410, 0.9417, 0.9405, 0.9404, 0.9404, 0.9408]
+
+plt.plot(X, tY)
+plt.plot(X, vY)
+plt.legend(["train_acc", "val_acc"])
+
+plt.show()

preprocessor.py

@@ -3,7 +3,7 @@ import nltk
 from nltk.corpus import stopwords
 from nltk.tokenize import word_tokenize
 from nltk.stem import WordNetLemmatizer
-#from spellchecker import SpellChecker as sc
+from spellchecker import SpellChecker as sc
 
 nltk.download("stopwords")
 nltk.download("punkt_tab")
@@ -31,3 +31,26 @@ def fr(r):
     #        r[k] = w
 
     return " ".join([l.lemmatize(i, pos = 'v') for i in r])
+
+def frs(r):
+    r = r.lower()
+
+    r = " ".join(tuple(re.findall(r'\w+', r)))
+
+    for i in ['\n', '\r', ',', '.', '-', ';', ':', '\'', '"']:
+        r = r.replace(i, "")
+
+    sw = set(stopwords.words("english"))
+    
+    l = WordNetLemmatizer()
+    c = sc()
+
+    r = [i.strip() for i in word_tokenize(r) if i.strip() not in sw]
+
+    # spellcheck
+    for k, i in enumerate(r):
+        w = c.correction(i)
+        if w:
+            r[k] = w
+
+    return " ".join([l.lemmatize(i, pos = 'v') for i in r])