1 ---
2 language: "id"
3 license: "mit"
4 datasets:
5 - wikipedia
6 widget:
7 - text: "Ibu ku sedang bekerja [MASK] sawah."
8 ---
9
10 # Indonesian BERT base model (uncased)
11
12 ## Model description
13 It is BERT-base model pre-trained with indonesian Wikipedia using a masked language modeling (MLM) objective. This
14 model is uncased: it does not make a difference between indonesia and Indonesia.
15
16 This is one of several other language models that have been pre-trained with indonesian datasets. More detail about
17 its usage on downstream tasks (text classification, text generation, etc) is available at [Transformer based Indonesian Language Models](https://github.com/cahya-wirawan/indonesian-language-models/tree/master/Transformers)
18
19 ## Intended uses & limitations
20
21 ### How to use
22 You can use this model directly with a pipeline for masked language modeling:
23 ```python
24 >>> from transformers import pipeline
25 >>> unmasker = pipeline('fill-mask', model='cahya/bert-base-indonesian-522M')
26 >>> unmasker("Ibu ku sedang bekerja [MASK] supermarket")
27
28 [{'sequence': '[CLS] ibu ku sedang bekerja di supermarket [SEP]',
29 'score': 0.7983310222625732,
30 'token': 1495},
31 {'sequence': '[CLS] ibu ku sedang bekerja. supermarket [SEP]',
32 'score': 0.090003103017807,
33 'token': 17},
34 {'sequence': '[CLS] ibu ku sedang bekerja sebagai supermarket [SEP]',
35 'score': 0.025469014421105385,
36 'token': 1600},
37 {'sequence': '[CLS] ibu ku sedang bekerja dengan supermarket [SEP]',
38 'score': 0.017966199666261673,
39 'token': 1555},
40 {'sequence': '[CLS] ibu ku sedang bekerja untuk supermarket [SEP]',
41 'score': 0.016971781849861145,
42 'token': 1572}]
43 ```
44 Here is how to use this model to get the features of a given text in PyTorch:
45 ```python
46 from transformers import BertTokenizer, BertModel
47
48 model_name='cahya/bert-base-indonesian-522M'
49 tokenizer = BertTokenizer.from_pretrained(model_name)
50 model = BertModel.from_pretrained(model_name)
51 text = "Silakan diganti dengan text apa saja."
52 encoded_input = tokenizer(text, return_tensors='pt')
53 output = model(**encoded_input)
54 ```
55 and in Tensorflow:
56 ```python
57 from transformers import BertTokenizer, TFBertModel
58
59 model_name='cahya/bert-base-indonesian-522M'
60 tokenizer = BertTokenizer.from_pretrained(model_name)
61 model = TFBertModel.from_pretrained(model_name)
62 text = "Silakan diganti dengan text apa saja."
63 encoded_input = tokenizer(text, return_tensors='tf')
64 output = model(encoded_input)
65 ```
66
67 ## Training data
68
69 This model was pre-trained with 522MB of indonesian Wikipedia.
70 The texts are lowercased and tokenized using WordPiece and a vocabulary size of 32,000. The inputs of the model are
71 then of the form:
72
73 ```[CLS] Sentence A [SEP] Sentence B [SEP]```
74