🏆 TurkReviewSentiment-RoBERTa: Sentiment Analysis Model for Turkish Texts

📢 A fine-tuned XLM-RoBERTa model for sentiment analysis in Turkish texts!
This model can classify user reviews and texts as positive or negative sentiment.

---

🔍 Model Details

Model Description

✅ Base Model: XLM-RoBERTa
✅ Finetuned from: xlm-roberta-base
✅ Training Data: User reviews from e-commerce platforms
✅ Task: Sentiment Analysis (Binary Classification: Positive / Negative)
✅ Language: Turkish
✅ Use Cases: Customer feedback analysis, social media monitoring, market research
✅ License: MIT

Model Sources

• Repository: Hugging Face Model Page

---

📌 How to Use the Model

import torch
from transformers import AutoModelForSequenceClassification, AutoTokenizer

def load_model(model_path):
    model = AutoModelForSequenceClassification.from_pretrained(model_path)
    tokenizer = AutoTokenizer.from_pretrained(model_path)
    return model, tokenizer

def get_sentiment(text, model, tokenizer):
    inputs = tokenizer(text, return_tensors="pt", truncation=True, padding=True, max_length=512)
    device = next(model.parameters()).device
    inputs = {k: v.to(device) for k, v in inputs.items()}
    outputs = model(**inputs)
    logits = outputs.logits
    predicted_class = torch.argmax(logits, dim=1).item()
    return "positive" if predicted_class == 1 else "negative"

    
model_path = "fundaylnci/TurkReviewSentiment-RoBERTa"
model, tokenizer = load_model(model_path)

text ="Bu ürün harika!"
get_sentiment(text,model, tokenizer)

## Training Details

### Training Data

Dataset: E-commerce product reviews + manually labeled sentiment data


#### Preprocessing [optional]

Tokenization with xlm-roberta-base tokenizer


#### Training Hyperparameters

        training_args = TrainingArguments(
            output_dir="./results",
            evaluation_strategy="epoch",
            learning_rate=learning_rate,
            per_device_train_batch_size=batch_size,
            num_train_epochs=3,
            weight_decay=0.01,
            logging_dir="./logs",
            logging_steps=10,
        )

#### Speeds, Sizes, Times [optional]

Checkpoint Size: ~500MB

Training Time: ~6 hours on NVIDIA T4 GPU

## Evaluation

Evaluation Metrics: Accuracy, F1-score

# Version 149608e489c0979a661489d53245a30a48d3d38e
Accuracy: 0.961
Precision: 0.977
Recall: 0.980
F1 Score: 0.979


## Environmental Impact

Carbon emissions can be estimated using the Machine Learning Impact calculator.

Hardware: NVIDIA T4 GPU
Training Hours: ~6 hours
Cloud Provider: Google Cloud
Estimated CO2 Emitted: ~3.2 kg

## Citation [optional]

**BibTeX:**

@article{turkreviewsentiment2025,
  title={TurkReviewSentiment-RoBERTa: Sentiment Analysis for Turkish Reviews},
  author={Your Name},
  journal={Hugging Face Model Hub},
  year={2025}
}