auskola/sentimientos

Sentiment Analysis Model

Model Details

• Base Model: google/electra-base-discriminator
• Task: Binary Sentiment Analysis (Positive/Negative)
• Datasets: IMDB and Amazon Reviews
• Language: English

Training Hyperparameters

• Batch Size: 8
• Learning Rate: 2e-5
• Number of Epochs: 2
• Max Sequence Length: 128 tokens
• Model Architecture: ELECTRA (Discriminator)

Training

The model was trained using a combination of IMDB and Amazon reviews datasets, using ELECTRA's discriminator architecture which is particularly efficient with limited data. The hyperparameters were optimized for performance on consumer-grade hardware.

Usage

from transformers import AutoModelForSequenceClassification, AutoTokenizer
import torch

# Load model and tokenizer
model_name = "auskola/sentimientos"
tokenizer = AutoTokenizer.from_pretrained(model_name)
model = AutoModelForSequenceClassification.from_pretrained(model_name)

def analyze_sentiment(text):
    # Tokenize and predict
    inputs = tokenizer(text, return_tensors="pt", truncation=True, max_length=128)
    with torch.no_grad():
        outputs = model(**inputs)
        probabilities = torch.nn.functional.softmax(outputs.logits, dim=1)
        
    # Get prediction and confidence
    prediction = torch.argmax(probabilities, dim=1)
    confidence = torch.max(probabilities).item()
    
    return {
        "sentiment": "Positive" if prediction.item() == 1 else "Negative",
        "confidence": confidence
    }

# Ejemplos de uso
texts = [
    "This product exceeded my expectations!",
    "Terrible service, would not recommend",
    "The movie was pretty good"
]

for text in texts:
    result = analyze_sentiment(text)
    print(f"\nText: {text}")
    print(f"Sentiment: {result['sentiment']}")
    print(f"Confidence: {result['confidence']:.2f}")