Update app.py

app.py

@@ -1,95 +1,95 @@
-import gradio as gr
-import torch
-from transformers import T5Tokenizer
-import torch.nn as nn
-from transformers import T5EncoderModel
-import re
-from nltk.tokenize import word_tokenize
-from nltk.corpus import stopwords
-from nltk.stem import WordNetLemmatizer
-import nltk
-
-# Download NLTK resources (only first time)
-nltk.download('punkt')
-nltk.download('stopwords')
-nltk.download('wordnet')
-nltk.download('omw-1.4')
-
-# Initialize preprocessing tools
-stop_words = set(stopwords.words('english'))
-lemmatizer = WordNetLemmatizer()
-
-# Preprocessing function
-def preprocess_text(text):
-    # Remove non-alphabet characters
-    text = re.sub(r'[^A-Za-z\s]', '', text)
-    # Remove URLs
-    text = re.sub(r'http\S+|www\S+|https\S+', '', text)
-    # Normalize whitespace
-    text = re.sub(r'\s+', ' ', text).strip()
-    # Lowercase
-    text = text.lower()
-    # Tokenize
-    tokens = word_tokenize(text)
-    # Remove stopwords
-    tokens = [word for word in tokens if word not in stop_words]
-    # Lemmatize
-    tokens = [lemmatizer.lemmatize(word) for word in tokens]
-    # Re-join
-    return ' '.join(tokens)
-
-# Model class
-class T5_regression(nn.Module):
-    def __init__(self):
-        super(T5_regression, self).__init__()
-        self.t5 = T5EncoderModel.from_pretrained("t5-base")
-        self.fc = nn.Linear(self.t5.config.d_model, 1)
-        self.relu = nn.ReLU()
-
-    def forward(self, input_ids, attention_mask):
-        output = self.t5(input_ids=input_ids, attention_mask=attention_mask)
-        pooled_output = output.last_hidden_state[:, 0, :]
-        rating = self.fc(pooled_output)
-        return rating.squeeze(-1)
-
-# Load tokenizer and model
-tokenizer = T5Tokenizer.from_pretrained("t5-base")
-device = torch.device("cuda" if torch.cuda.is_available() else "cpu")
-model = T5_regression().to(device)
-
-# Load trained weights
-model.load_state_dict(torch.load("best_model.pth", map_location=device))
-model.eval()
-
-# Prediction function
-def predict_rating(review_text):
-    # Preprocess review
-    clean_text = preprocess_text(review_text)
-
-    encoding = tokenizer(
-        clean_text,
-        truncation=True,
-        padding='max_length',
-        max_length=128,
-        return_tensors='pt'
-    )
-
-    input_ids = encoding['input_ids'].to(device)
-    attention_mask = encoding['attention_mask'].to(device)
-
-    with torch.no_grad():
-        output = model(input_ids, attention_mask)
-        rating = output.item()
-
-    return round(rating, 1)
-
-# Gradio UI
-iface = gr.Interface(
-    fn=predict_rating,
-    inputs=gr.Textbox(lines=4, placeholder="Enter your review here..."),
-    outputs=gr.Number(label="Predicted Rating"),
-    title="Review Rating Predictor",
-    description="Predicts the rating of a mobile app review using a fine-tuned T5 regression model."
-)
-
-iface.launch()
+import gradio as gr
+import torch
+from transformers import T5Tokenizer
+import torch.nn as nn
+from transformers import T5EncoderModel
+import re
+from nltk.tokenize import word_tokenize
+from nltk.corpus import stopwords
+from nltk.stem import WordNetLemmatizer
+import nltk
+
+# Download NLTK resources (only first time)
+nltk.download('punkt_tab')
+nltk.download('stopwords')
+nltk.download('wordnet')
+nltk.download('omw-1.4')
+
+# Initialize preprocessing tools
+stop_words = set(stopwords.words('english'))
+lemmatizer = WordNetLemmatizer()
+
+# Preprocessing function
+def preprocess_text(text):
+    # Remove non-alphabet characters
+    text = re.sub(r'[^A-Za-z\s]', '', text)
+    # Remove URLs
+    text = re.sub(r'http\S+|www\S+|https\S+', '', text)
+    # Normalize whitespace
+    text = re.sub(r'\s+', ' ', text).strip()
+    # Lowercase
+    text = text.lower()
+    # Tokenize
+    tokens = word_tokenize(text)
+    # Remove stopwords
+    tokens = [word for word in tokens if word not in stop_words]
+    # Lemmatize
+    tokens = [lemmatizer.lemmatize(word) for word in tokens]
+    # Re-join
+    return ' '.join(tokens)
+
+# Model class
+class T5_regression(nn.Module):
+    def __init__(self):
+        super(T5_regression, self).__init__()
+        self.t5 = T5EncoderModel.from_pretrained("t5-base")
+        self.fc = nn.Linear(self.t5.config.d_model, 1)
+        self.relu = nn.ReLU()
+
+    def forward(self, input_ids, attention_mask):
+        output = self.t5(input_ids=input_ids, attention_mask=attention_mask)
+        pooled_output = output.last_hidden_state[:, 0, :]
+        rating = self.fc(pooled_output)
+        return rating.squeeze(-1)
+
+# Load tokenizer and model
+tokenizer = T5Tokenizer.from_pretrained("t5-base")
+device = torch.device("cuda" if torch.cuda.is_available() else "cpu")
+model = T5_regression().to(device)
+
+# Load trained weights
+model.load_state_dict(torch.load("best_model.pth", map_location=device))
+model.eval()
+
+# Prediction function
+def predict_rating(review_text):
+    # Preprocess review
+    clean_text = preprocess_text(review_text)
+
+    encoding = tokenizer(
+        clean_text,
+        truncation=True,
+        padding='max_length',
+        max_length=128,
+        return_tensors='pt'
+    )
+
+    input_ids = encoding['input_ids'].to(device)
+    attention_mask = encoding['attention_mask'].to(device)
+
+    with torch.no_grad():
+        output = model(input_ids, attention_mask)
+        rating = output.item()
+
+    return round(rating, 1)
+
+# Gradio UI
+iface = gr.Interface(
+    fn=predict_rating,
+    inputs=gr.Textbox(lines=4, placeholder="Enter your review here..."),
+    outputs=gr.Number(label="Predicted Rating"),
+    title="Review Rating Predictor",
+    description="Predicts the rating of a mobile app review using a fine-tuned T5 regression model."
+)
+
+iface.launch()