Ha1dir/sentimen-indobert

IndoBERT Sentiment Analysis Model

Model ini adalah hasil fine-tuning model IndoBERT base untuk tugas klasifikasi sentimen bahasa Indonesia.

Penggunaan

Use a pipeline as a high-level helper

from transformers import pipeline pipe = pipeline("text-classification", model="Ha1dir/sentimen-indobert")

Load model directly

from transformers import AutoTokenizer, AutoModelForSequenceClassification tokenizer = AutoTokenizer.from_pretrained("Ha1dir/sentimen-indobert") model = AutoModelForSequenceClassification.from_pretrained("Ha1dir/sentimen-indobert")

Label Kelas

• 0: Positive
• 1: Negative
• 2: Neutral

Tentang Model

• Base Model: indobenchmark/indobert-base-p1
• Training Epochs: 5
• Optimizer: Adam, LR = 3e-6

Hasil Training

(Epoch 1) TRAIN LOSS: 0.2962 Acc: 0.8896 Precision: 0.8893 Recall: 0.8896 F1: 0.8876

(Epoch 2) TRAIN LOSS: 0.1450 Acc: 0.9514 Precision: 0.9513 Recall: 0.9514 F1: 0.9513

(Epoch 3) TRAIN LOSS: 0.1053 Acc: 0.9646 Precision: 0.9646 Recall: 0.9646 F1: 0.9646

(Epoch 4) TRAIN LOSS: 0.0722 Acc: 0.9781 Precision: 0.9781 Recall: 0.9781 F1: 0.9781

(Epoch 5) TRAIN LOSS: 0.0468 Acc: 0.9874 Precision: 0.9874 Recall: 0.9874 F1: 0.9874

Validasi Mode

VAL LOSS: 0.0234 Acc: 0.9955 Precision: 0.9955 Recall: 0.9955 F1: 0.9954

Evaluasi

VAL LOSS: 0.0234 Acc: 0.9982 Precision: 0.9982 Recall: 0.9982 F1: 0.9982

---