IndoQA

This model is a fine-tuned version of indolem/indobert-base-uncased on jakartaresearch/indoqa. It achieves the following results on the evaluation set:

• Loss: 1.4807

Training hyperparameters

The following hyperparameters were used during training:

• learning_rate: 2e-05
• train_batch_size: 16
• eval_batch_size: 16
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 3

Training results

Training Loss	Epoch	Step	Validation Loss
No log	1.0	207	1.9698
No log	2.0	414	1.8862
0.9416	3.0	621	1.4807

How to use this model in Transformers Library

from transformers import pipeline

question = "Berapa jumlah pulau yang ada di indonesia?"
context = "Indonesia adalah negara kepulauan, Dengan jumlah pulau sekitar 17 ribu"

from transformers import pipeline

question_answerer = pipeline("question-answering", model="digo-prayudha/IndoQA")
question_answerer(question=question, context=context)

Framework versions

• Transformers 4.35.2
• Pytorch 2.1.0+cu121
• Datasets 2.15.0
• Tokenizers 0.15.0