Spaces:
Sleeping
Sleeping
| import streamlit as st | |
| from transformers import AutoTokenizer, TFAutoModelForSequenceClassification | |
| import tensorflow as tf | |
| # Constants | |
| MODEL_NAME = "distilbert-base-uncased-finetuned-sst-2-english" # Pre-trained model for sentiment analysis | |
| # Load the tokenizer and model | |
| def load_model(): | |
| tokenizer = AutoTokenizer.from_pretrained(MODEL_NAME) | |
| model = TFAutoModelForSequenceClassification.from_pretrained(MODEL_NAME) | |
| return tokenizer, model | |
| # Function to classify text | |
| def classify_text(input_text, tokenizer, model): | |
| inputs = tokenizer(input_text, return_tensors="tf", truncation=True, padding=True) | |
| outputs = model(inputs) | |
| logits = outputs.logits | |
| probabilities = tf.nn.softmax(logits, axis=-1) | |
| return probabilities.numpy() | |
| # Streamlit UI | |
| def main(): | |
| st.title("Hugging Face + TensorFlow: Sentiment Analysis Demo") | |
| st.write("Analyze the sentiment of a given text using a Hugging Face pre-trained model with TensorFlow.") | |
| # User input | |
| user_input = st.text_area("Enter your text below:", height=150) | |
| if st.button("Analyze"): | |
| if user_input.strip(): | |
| with st.spinner("Analyzing sentiment..."): | |
| tokenizer, model = load_model() | |
| probabilities = classify_text(user_input, tokenizer, model) | |
| st.success("Analysis Complete!") | |
| # Display probabilities | |
| labels = ["Negative", "Positive"] | |
| results = {labels[i]: float(probabilities[0][i]) for i in range(len(labels))} | |
| st.write("### Sentiment Probabilities") | |
| st.json(results) | |
| # Display prediction | |
| predicted_label = labels[tf.argmax(probabilities[0])] | |
| st.write(f"### Predicted Sentiment: `{predicted_label}`") | |
| else: | |
| st.warning("Please enter some text to analyze.") | |
| # Footer | |
| st.markdown("---") | |
| st.markdown("**Built with Hugging Face Transformers and TensorFlow.**") | |
| if __name__ == "__main__": | |
| main() |