README.md

---
license: cc-by-3.0
datasets:
- stanfordnlp/imdb
language:
- en
metrics:
- accuracy
tags:
- sentiment
- small
- silly
- RETON
---

model structure:


Input Text

   |
   
Custom Tokenizer (Text -> Tokens)

   |
   
Padding (Equal Length Sequences)

   |
   
Embedding Layer (Tokens -> Dense Vectors)

   |
   
LSTM Layer (Capture Sequential Patterns)

   |
   
Attention Mechanism (Focus on Important Tokens)

   |
   
Dense Layer (Fully Connected)

   |
   
Softmax Activation (Output: Positive or Negative Sentiment




it archieved 95.71% accuracy on the dataset provided

to test it use smth like:


```python
import tensorflow as tf
from tensorflow.keras.preprocessing.sequence import pad_sequences
from tensorflow.keras.datasets import imdb


MAX_VOCAB_SIZE = 10000
MAX_SEQUENCE_LENGTH = 200
model = tf.keras.models.load_model('RETO-SENTIMENT.h5', custom_objects={'AttentionLayer': AttentionLayer})

class AttentionLayer(tf.keras.layers.Layer):
    def __init__(self, **kwargs):  
        super(AttentionLayer, self).__init__(**kwargs)

    def call(self, inputs):
        attention_weights = tf.nn.softmax(tf.reduce_sum(inputs, axis=-1), axis=-1)
        attention_weights = tf.expand_dims(attention_weights, -1)
        weighted_sum = inputs * attention_weights
        return tf.reduce_sum(weighted_sum, axis=1)

word_index = imdb.get_word_index()
reverse_word_index = {value: key for key, value in word_index.items()}


def preprocess_text(input_text):
    tokens = [word_index.get(word.lower(), 2) for word in input_text.split()]  # 2 is unknown
    padded_sequence = pad_sequences([tokens], maxlen=MAX_SEQUENCE_LENGTH)
    return padded_sequence


def predict_sentiment(input_text):
    preprocessed_text = preprocess_text(input_text)
    prediction = model.predict(preprocessed_text)
    sentiment = "Positive" if prediction[0][0] > 0.5 else "Negative"
    confidence = prediction[0][0] if sentiment == "Positive" else 1 - prediction[0][0]
    return sentiment, confidence


if __name__ == "__main__":
    test_sentences = [
#add sentences here
    ]

    for sentence in test_sentences:
        sentiment, confidence = predict_sentiment(sentence)
        print(f"Input: {sentence}")
        print(f"Predicted Sentiment: {sentiment} (Confidence: {confidence:.2f})\n")
```