Rubert Tiny2 Russian Emotion Sentiment

🚀 rubert_tiny2_russian_emotion_sentiment

The rubert_tiny2_russian_emotion_sentiment model is a fine - tuned version of the lightweight model. It can classify five emotions in Russian messages, offering a practical solution for sentiment analysis in the Russian language.

🚀 Quick Start

To start using the rubert_tiny2_russian_emotion_sentiment model, you first need to install the necessary dependencies:

pip install transformers torch

Then you can use the following code to perform emotion classification:

from transformers import AutoTokenizer, AutoModelForSequenceClassification
import torch

# Load the model and tokenizer
MODEL_ID = "Kostya165/rubert_tiny2_russian_emotion_sentiment"
tokenizer = AutoTokenizer.from_pretrained(MODEL_ID)
model     = AutoModelForSequenceClassification.from_pretrained(MODEL_ID)
model.eval()

texts = [
    "Сегодня отличный день!",
    "Меня это всё бесит и раздражает."
]

# Tokenization
enc = tokenizer(texts, padding=True, truncation=True, max_length=128, return_tensors="pt")
with torch.no_grad():
    logits = model(**enc).logits
    preds = logits.argmax(dim=-1).tolist()

# Convert IDs back to labels
id2label = model.config.id2label
labels = [id2label[p] for p in preds]
print(labels)  # e.g.: ['positive', 'aggression']

✨ Features

• Emotion Classification: Classify five emotions in Russian messages: aggression, anxiety, neutral, positive, and sarcasm.
• High Accuracy: Achieved high accuracy on the validation set.

📦 Installation

You need to install the following dependencies:

• transformers>=4.30.0
• torch>=1.10.0
• datasets
• evaluate

📚 Documentation

Validation Results

Per - class accuracy:

• aggression: 0.9120
• anxiety: 0.9462
• neutral: 0.8663
• positive: 0.8884
• sarcasm: 0.8426

Training Details

• Base Model: cointegrated/rubert-tiny2
• Dataset: Kostya165/ru_emotion_dvach
• Epochs: 2
• Batch size: 32
• Learning rate: 1e - 5
• Mixed precision: FP16
• Regularization: Dropout 0.1, weight_decay 0.01, warmup_ratio 0.1

📄 License

This model is licensed under CC - BY - SA 4.0.

📚 Citation

@article{rubert_tiny2_russian_emotion_sentiment,
  title   = {Russian Emotion Sentiment Classification with RuBERT-tiny2},
  author  = {Kostya165},
  year    = {2024},
  howpublished = {\url{https://huggingface.co/Kostya165/rubert_tiny2_russian_emotion_sentiment}}
}