Modelpublic

AleksandarDimitrov06/bg-ecommerce-sentiment-roberta

sourceHugging Facemitupdated 3mo agoView on Hugging Face

0likes5downloads

Model Card

RoBERTa Base Sentiment BG - Fine-tuned

This model is a fine-tuned version of `rmihaylov/roberta-base-sentiment-bg` for binary sentiment analysis (Positive/Negative). It has been trained on custom review data to classify text sentiment in Bulgarian and English.

Model Description

—Developed by: Aleksandar Dimitrov
—Language(s): Bulgarian (bg), English (en)
—Model Type: Text Classification (Sentiment Analysis)
—Base Model: rmihaylov/roberta-base-sentiment-bg
—License: MIT

Как да използвате модела (How to Use)

Можете да използвате модела директно с библиотеката transformers в Python. Ето пример как да заредите модела и да направите предсказание:

You can use this model directly with the transformers library in Python. Here is a script to load the model and run inference:

python

import torch
from transformers import AutoTokenizer, AutoModelForSequenceClassification

model_id = "AleksandarDimitrov06/bg-ecommerce-sentiment-roberta"

device = torch.device("cuda" if torch.cuda.is_available() else "cpu")
tokenizer = AutoTokenizer.from_pretrained(model_id)
model = AutoModelForSequenceClassification.from_pretrained(model_id).to(device)
model.eval()

def predict_sentiment(text: str):
    with torch.no_grad():
        inputs = tokenizer(text, return_tensors="pt", truncation=True, padding=True).to(device)
        outputs = model(**inputs)
        probabilities = torch.softmax(outputs.logits, dim=1).tolist()[0]
        
        # 0: NEGATIVE, 1: POSITIVE
        if probabilities[1] >= probabilities[0]:
            return "POSITIVE", probabilities[1]
        return "NEGATIVE", probabilities[0]

# Тест / Test
text = "Този продукт е страхотен и работи перфектно!"
label, confidence = predict_sentiment(text)
print(f"Резултат: {label} ({confidence:.2%})")