Modelpublic

msamilim/modernbert-turkish-sentiment-optuna-hpo

sourceHugging Faceapache-2.0updated 10mo agoView on Hugging Face

0likes14downloads

Model Card

Turkish Sentiment Analysis (3-class) — Fine-tuned

Overview

This model is a fine-tuned version of `artiwise-ai/modernbert-base-tr-uncased` for 3-class Turkish sentiment analysis. It was trained on an imbalanced dataset of e-commerce product reviews, and hyperparameters were optimized with Optuna to obtain the most effective fine-tuning configuration.

Bu model, üç sınıflı Türkçe duygu analizi için `artiwise-ai/modernbert-base-tr-uncased` taban alınarak ince ayar (fine-tuning) yapılmış bir sürümdür. Model, dengesiz bir e-ticaret ürün yorumları veri kümesi üzerinde eğitilmiş; en etkili ince ayar yapılandırmasını elde etmek için hiperparametreler Optuna ile optimize edilmiştir.

Intended Use

—Product reviews classification
—Social media analysis
—Customer feedback analysis
—Brand monitoring
—Market research
—Customer service optimization
—Competitive intelligence

Model Details

Field	Value
Model Name	msamilim/modernbert-turkish-sentiment-optuna-hpo
Base Model	artiwise-ai/modernbert-base-tr-uncased
Task	Sentiment Analysis
Language	Turkish
Fine-Tuning Dataset	Turkish E-Commerce Product Reviews Dataset
Number of Labels	3
Problem Type	Single-label classification
License	apache-2.0
Fine-Tuning Framework	Hugging Face Transformers

Dataset

The dataset is a Turkish three-class sentiment corpus (negatif / notr / pozitif). Overall distribution and per-split distributions are shown below.

Dataset Distribution (Overall)

LabelID	LabelName	Count	Ratio (%)
0	negatif	9462	18.86
1	notr	746	1.49
2	pozitif	39952	79.65
—	Total	50160	100.00

Training Procedure

—Objective metric: `eval_macro_f1`
—Hyperparameter Optimization Techniques: `Optuna`

HPO Parameter Ranges

json

params = {
        "learning_rate": trial.suggest_float("learning_rate", 5e-6, 5e-5, log=True),
        "per_device_train_batch_size": trial.suggest_categorical("per_device_train_batch_size", [16, 32]),
        "per_device_eval_batch_size":  trial.suggest_categorical("per_device_eval_batch_size",  [32]),
        "weight_decay": trial.suggest_float("weight_decay", 0.0, 0.1),
        "warmup_ratio": trial.suggest_float("warmup_ratio", 0.0, 0.2),
        "num_train_epochs": trial.suggest_int("num_train_epochs", 6, 8),
        "gradient_accumulation_steps": trial.suggest_categorical("gradient_accumulation_steps", [1, 2, 4]),
    }

Best Trial Hyperparameters

json

{
  "learning_rate": 4.292071084445416e-05,
  "per_device_train_batch_size": 16,
  "per_device_eval_batch_size": 32,
  "weight_decay": 0.0897155306921003,
  "warmup_ratio": 0.0495607080384387,
  "num_train_epochs": 7,
  "gradient_accumulation_steps": 2
}

Evaluation Results

These results are the evaluations recorded during the final fine-tuning training process. | label | precision | recall | f1-score | | --- | --- | --- | --- | | negatif | 0.8467 | 0.7930 | 0.8189 | | notr | 0.1238 | 0.1444 | 0.1333 | | pozitif | 0.9555 | 0.9668 | 0.9611 | | accuracy | | | 0.9218 | | micro avg | 0.9218 | 0.9218 | 0.9218 | | macro avg | 0.6420 | 0.6347 | 0.6378 | | weighted avg | 0.9225 | 0.9218 | 0.9219 |

Epoch-wise Metrics

epoch	train_loss	eval_loss	eval_macro_f1
1	0.5036	0.2085	0.5909
2	0.2767	0.2371	0.6065
3	0.1249	0.4557	0.6378
4	0.0349	0.6602	0.6347
5	0.0099	0.7144	0.6217
6	0.0025	0.7635	0.6341
7	0.0002	0.7692	0.6322

How to use - Pipeline

python

from transformers import pipeline

# Load the classification pipeline with the specified model
model_name = "msamilim/modernbert-turkish-sentiment-optuna-hpo"
pipe = pipeline("text-classification", model=model_name)

# Classify a new sentence
sentence = "Güzel ürün, tavsiye ederim."
result = pipe(sentence)

# Print the result
print(result)

# Example output : 
# [{'label': 'pozitif', 'score': 0.9998408555984497}]

How to use - Full Classification

python

from transformers import AutoTokenizer, AutoModelForSequenceClassification
import torch

model_name = "msamilim/modernbert-turkish-sentiment-optuna-hpo"
tokenizer = AutoTokenizer.from_pretrained(model_name)
model = AutoModelForSequenceClassification.from_pretrained(model_name)

def predict_sentiment(texts):
    inputs = tokenizer(texts, return_tensors="pt", truncation=True, padding=True, max_length=512)
    with torch.no_grad():
        outputs = model(**inputs)
    probabilities = torch.nn.functional.softmax(outputs.logits, dim=-1)
    id2label = {    0: "Negatif",    1: "Nötr",    2: "Pozitif"}
    return [id2label[p] for p in torch.argmax(probabilities, dim=-1).tolist()]

texts = [
     "Güzel ürün, tavsiye ederim kullanılır.", 
     "Ürün çok güzel ve kaliteli. Maalesef yüzüme uymadığı için iade etmek zorunda kaldım.", 
     "Keşke aldıktan sonra indirime girmeseydi.",
     "Daha soluk ve mat yapısı var beğenmedim .",
]

for text, sentiment in zip(texts, predict_sentiment(texts)):
    print(f"Text: {text}\nSentiment: {sentiment}\n")

# Example output : 
# Text: Güzel ürün, tavsiye ederim kullanılır.
# Sentiment: Pozitif
# Text: Ürün çok güzel ve kaliteli. Maalesef yüzüme uymadığı için iade etmek zorunda kaldım.
# Sentiment: Pozitif
# Text: Keşke aldıktan sonra indirime girmeseydi.
# Sentiment: Negatif
# Text: Daha soluk ve mat yapısı var beğenmedim .
# Sentiment: Negatif

Framework versions

—transformers==4.57.0
—torch==2.8.0+cu128
—datasets==4.2.0
—accelerate==1.10.1
—evaluate==0.4.6
—python==3.11.13