MS to IND HuggingFace#

This tutorial is available as an IPython notebook at Malaya/example/ms-ind-translation-huggingface.

This module trained on standard language and augmented local language structures, proceed with caution.

[1]:

import os

os.environ['CUDA_VISIBLE_DEVICES'] = ''

[2]:

%%time

import malaya
import logging

logging.basicConfig(level=logging.INFO)

CPU times: user 3.81 s, sys: 3.51 s, total: 7.32 s
Wall time: 3.19 s

List available HuggingFace models#

[3]:

malaya.translation.ms_ind.available_huggingface()

INFO:malaya.translation.ms_ind:tested on FLORES200 MS-IND (zsm_Latn-ind_Latn) pair `dev` set, https://github.com/facebookresearch/flores/tree/main/flores200

[3]:
Size (MB) BLEU SacreBLEU Verbose SacreBLEU-chrF++-FLORES200 Suggested length
mesolitica/finetune-translation-austronesian-t5-tiny-standard-bahasa-cased 139 33.882077 67.7/42.2/28.0/18.9 (BP = 0.966 ratio = 0.966 ... 59.46 512
mesolitica/finetune-translation-austronesian-t5-small-standard-bahasa-cased 242 35.954481 66.3/42.6/29.1/20.3 (BP = 1.000 ratio = 1.014 ... 61.02 512
mesolitica/finetune-translation-austronesian-t5-base-standard-bahasa-cased 892 37.62068 70.0/45.8/31.7/22.5 (BP = 0.967 ratio = 0.968 ... 62.1 512

Load Transformer models#

def huggingface(
    model: str = 'mesolitica/finetune-translation-austronesian-t5-small-standard-bahasa-cased',
    force_check: bool = True,
    **kwargs,
):
    """
    Load HuggingFace model to translate MS-to-IND.

    Parameters
    ----------
    model: str, optional (default='mesolitica/finetune-translation-t5-small-standard-bahasa-cased')
        Check available models at `malaya.translation.ms_ind.available_huggingface()`.

    Returns
    -------
    result: malaya.torch_model.huggingface.Generator
    """

[4]:

transformer_huggingface = malaya.translation.ms_ind.huggingface()

Translate#

def generate(self, strings: List[str], **kwargs):
    """
    Generate texts from the input.

    Parameters
    ----------
    strings : List[str]
    **kwargs: vector arguments pass to huggingface `generate` method.
        Read more at https://huggingface.co/docs/transformers/main_classes/text_generation

    Returns
    -------
    result: List[str]
    """

For better results, always split by end of sentences.

[5]:

from pprint import pprint

[8]:

news = "SHAH ALAM - Kerajaan harus bersikap 'kejam' terhadap desakan segelintir rakyat berkenaan isu pengeluaran Kumpulan Wang Simpanan Pekerja (KWSP) demi kesejahteraan rakyat di masa akan datang."

[9]:

%%time

pprint(transformer_huggingface.generate([news],
                                       max_length = 1000))

["SHAH ALAM - Pemerintah harus 'kejam' dengan desakan segelintir masyarakat "
 'terkait isu penarikan Dana Penyediaan Pegawai (EPF) untuk kesejahteraan '
 'masyarakat di masa depan.']
CPU times: user 3.53 s, sys: 0 ns, total: 3.53 s
Wall time: 298 ms

compare with Google translate using googletrans#

Install it by,

pip3 install googletrans==4.0.0rc1

[10]:

from googletrans import Translator

translator = Translator()

[11]:

strings = [news]

[12]:

for t in strings:
    r = translator.translate(t, src='ms', dest = 'id')
    print(r.text)

Shah Alam - Pemerintah harus 'kejam' terhadap desakan beberapa orang tentang masalah produksi Dana Penyedia Karyawan (EPF) untuk masa depan rakyat.