Indonesian tiny-RoBERTa
How to Use
As Masked Language Model
from transformers import pipeline
pretrained_name = "akahana/tiny-roberta-indonesia"
fill_mask = pipeline(
"fill-mask",
model=pretrained_name,
tokenizer=pretrained_name
)
fill_mask("ikiryo adalah <mask> hantu dalam mitologi jepang.")
Feature Extraction in PyTorch
from transformers import RobertaModel, RobertaTokenizerFast
pretrained_name = "akahana/tiny-roberta-indonesia"
model = RobertaModel.from_pretrained(pretrained_name)
tokenizer = RobertaTokenizerFast.from_pretrained(pretrained_name)
prompt = "ikiryo adalah <mask> hantu dalam mitologi jepang."
encoded_input = tokenizer(prompt, return_tensors='pt')
output = model(**encoded_input)
