Update app.py

app.py

@@ -1,44 +1,3 @@
-import tensorflow as tf
-from tensorflow import keras
-from tensorflow.keras.layers import TextVectorization, Embedding, GlobalAveragePooling1D, Dense
-import pandas as pd
-from sklearn.model_selection import train_test_split
-from datasets import load_dataset
-import gradio as gr
-
-# Load dataset
-ds = load_dataset("uhoui/text-tone-classifier")
-df = pd.DataFrame(ds['train'])
-
-# Remove classes with only one sample
-class_counts = df['label'].value_counts()
-classes_to_keep = class_counts[class_counts > 1].index
-df = df[df['label'].isin(classes_to_keep)]
-
-# Convert labels to categorical
-df['label'] = pd.Categorical(df['label'])
-df['label'] = df['label'].cat.codes  # Convert to numerical codes
-
-# Split data without stratification
-train_texts, test_texts, train_labels, test_labels = train_test_split(
-    df['text'], df['label'], test_size=0.2, random_state=42
-)
-
-# Convert labels to NumPy arrays for TensorFlow
-train_labels = train_labels.to_numpy()
-test_labels = test_labels.to_numpy()
-
-# Convert Pandas Series to TensorFlow datasets
-train_ds = tf.data.Dataset.from_tensor_slices((train_texts.values, train_labels))
-test_ds = tf.data.Dataset.from_tensor_slices((test_texts.values, test_labels))
-
-# Text Vectorization
-vocab_size = 10000
-vectorizer = TextVectorization(max_tokens=vocab_size, output_mode="int", output_sequence_length=50)
-vectorizer.adapt(train_texts.to_numpy())
-
-# Build the model
-def build_model():
     model = keras.Sequential([
         vectorizer,
         Embedding(input_dim=vocab_size, output_dim=64),
@@ -49,22 +8,18 @@ def build_model():
     
     model.compile(optimizer='adam', loss='sparse_categorical_crossentropy', metrics=['accuracy'])
     return model
-
 # Train the model
 model = build_model()
 model.fit(train_ds.batch(32), epochs=10, validation_data=test_ds.batch(32))
-
 # Function to make predictions
 def predict(text):
     # Vectorize the input text
     vectorized_text = vectorizer([text])  # Use the vectorizer to transform the input
-    prediction = model.predict(vectorized_text)  # No need to expand dimensions
+    prediction = model.predict(vectorized_text)  # Pass the vectorized input to the model
     predicted_label = tf.argmax(prediction, axis=1).numpy()[0]
     return df['label'].cat.categories[predicted_label]
-
 # Gradio interface
 iface = gr.Interface(fn=predict, inputs="text", outputs="text", title="Text Tone Sentiment Analysis",
                      description="Enter a text to analyze its tone (e.g., joy, depression, contentment).")
-
 if __name__ == "__main__":
-    iface.launch()
+    iface.launch()