Bahasa ELECTRA Model

Pretrained ELECTRA small language model for Malay and Indonesian.

Pretraining Corpus

electra-small-generator-bahasa-cased model was pretrained on ~1.8 Billion words. We trained on both standard and social media language structures, and below is list of data we trained on,

  1. dumping wikipedia.
  2. local instagram.
  3. local twitter.
  4. local news.
  5. local parliament text.
  6. local singlish/manglish text.
  7. IIUM Confession.
  8. Wattpad.
  9. Academia PDF.

Preprocessing steps can reproduce from here, Malaya/pretrained-model/preprocess.

Pretraining details

Load Pretrained Model

You can use this model by installing torch or tensorflow and Huggingface library transformers. And you can use it directly by initializing it like this:

from transformers import ElectraTokenizer, ElectraModel

model = ElectraModel.from_pretrained('huseinzol05/electra-small-generator-bahasa-cased')
tokenizer = ElectraTokenizer.from_pretrained(
    do_lower_case = False,

Example using AutoModelWithLMHead

from transformers import ElectraTokenizer, AutoModelWithLMHead, pipeline

model = AutoModelWithLMHead.from_pretrained('huseinzol05/electra-small-generator-bahasa-cased')
tokenizer = ElectraTokenizer.from_pretrained(
    do_lower_case = False,
fill_mask = pipeline('fill-mask', model = model, tokenizer = tokenizer)
print(fill_mask('makan ayam dengan [MASK]'))

Output is,

[{'sequence': '[CLS] makan ayam dengan ayam [SEP]',
  'score': 0.08424834907054901,
  'token': 3255},
 {'sequence': '[CLS] makan ayam dengan rendang [SEP]',
  'score': 0.064150370657444,
  'token': 6288},
 {'sequence': '[CLS] makan ayam dengan nasi [SEP]',
  'score': 0.033446669578552246,
  'token': 2533},
 {'sequence': '[CLS] makan ayam dengan kucing [SEP]',
  'score': 0.02803465723991394,
  'token': 3577},
 {'sequence': '[CLS] makan ayam dengan telur [SEP]',
  'score': 0.026627106592059135,
  'token': 6350}]


For further details on the model performance, simply checkout accuracy page from Malaya, https://malaya.readthedocs.io/en/latest/Accuracy.html, we compared with traditional models.


Thanks to Im Big, LigBlou, Mesolitica and KeyReply for sponsoring AWS, Google and GPU clouds to train ELECTRA for Bahasa.

Downloads last month
Hosted inference API
Mask token: [MASK]
This model can be loaded on the Inference API on-demand.