File size: 12,745 Bytes
fc82b7f
b93f6a6
fc82b7f
b93f6a6
d9d9313
 
b93f6a6
 
d9d9313
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
fc82b7f
b93f6a6
 
 
e159f47
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
13f46dc
e159f47
 
 
 
 
 
 
 
 
 
 
 
13f46dc
e159f47
 
 
 
 
 
 
13f46dc
e159f47
 
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
---
language: id
license: mit
datasets:
- indonli
- MoritzLaurer/multilingual-NLI-26lang-2mil7
pipeline_tag: zero-shot-classification
widget:
- text: Saya suka makan kentang goreng.
  candidate_labels: positif, netral, negatif
  hypothesis_template: Kalimat ini mengandung tema {}.
  multi_class: false
  example_title: Sentiment
- text: Apple umumkan harga iPhone 14.
  candidate_labels: teknologi, olahraga, kuliner, bisnis
  hypothesis_template: Kalimat ini mengandung tema {}.
  multi_class: true
  example_title: News
model-index:
- name: ilos-vigil/bigbird-small-indonesian-nli
  results:
  - task:
      type: natural-language-inference
      name: Natural Language Inference
    dataset:
      name: indonli
      type: indonli
      config: indonli
      split: test_expert
    metrics:
    - type: accuracy
      value: 0.5385388739946381
      name: Accuracy
      verified: true
      verifyToken: eyJhbGciOiJFZERTQSIsInR5cCI6IkpXVCJ9.eyJoYXNoIjoiNWRhZDkxNmI2NzE3MzRlYmNlMWFjZDVmNWUwYmMwN2IxYzNjMWE4YzY4NWI3NDZkYTMzY2NjN2MyZGQ5YzEwZSIsInZlcnNpb24iOjF9.AgizskHeXOzs0v93DNojNoqR_-1bQsYBokL8jcfelFm-zt-r5YXt89WXBDLLg4oKv-Roj8sLhUwe7ei0Mf1-Ag
    - type: f1
      value: 0.530444188199697
      name: F1 Macro
      verified: true
      verifyToken: eyJhbGciOiJFZERTQSIsInR5cCI6IkpXVCJ9.eyJoYXNoIjoiMjk2YTFhY2E3NGIzNzgxY2M5YzUzNGUzYTAwOWZkNGU3Y2I5MDA1MTc0YzM4Yjg0MmIzY2Y5M2EzOGYxNjY4NiIsInZlcnNpb24iOjF9.YZ_fTuVftTCM6SFfkFCLPbJWYmYNMYL9PNHUwNFHQXZeknf6OCBgQtr1gF6VM9mX6WuU4OKEl12tsAytlkm7Ag
    - type: f1
      value: 0.5385388739946381
      name: F1 Micro
      verified: true
      verifyToken: eyJhbGciOiJFZERTQSIsInR5cCI6IkpXVCJ9.eyJoYXNoIjoiM2MxMGUyZmJhZTYzN2M4NDlkMTZmMzllOGVhMjRiODhkMGVkMGMxMjY2NDBkZWM3ZWY2ZjhmZTNmYWU5ZjEzMyIsInZlcnNpb24iOjF9.f0HQlPRx4VFnOOHsrvMKFni8g1B1OJfheOyADsf47GnrvCcW_dakDgBy5c_yy4TehQYRa6ToYGHnuQnemvhnBg
    - type: f1
      value: 0.5299257731385174
      name: F1 Weighted
      verified: true
      verifyToken: eyJhbGciOiJFZERTQSIsInR5cCI6IkpXVCJ9.eyJoYXNoIjoiNTgzZjJkZWU0NDgyMGU5MDFmNzk2OWY1OWY4MzA2NTE3MDAxN2Y2MWExODJkYjdlN2I1YzgzYjljNjdkMTc1YiIsInZlcnNpb24iOjF9.lWB7MZlAiDjskKM-lx-XtLxTQYuWLz3QjyseDuZe_AxtyOKt2GZkP2NDOZxEWketHjRiTCQfBUvSfzFId-FCAg
    - type: precision
      value: 0.5592571894118881
      name: Precision Macro
      verified: true
      verifyToken: eyJhbGciOiJFZERTQSIsInR5cCI6IkpXVCJ9.eyJoYXNoIjoiZDQxYTFlNTNjNDAwMWIxYmJlMzRkN2U5OWY1NWNjN2YyYTE2NzRjNjM3ZWNhMzM4NjFhYWM4MzJkYjY3MzU0YSIsInZlcnNpb24iOjF9.6OI4_M1wLX1Z1BztKUfZ-382F3coCeJjarsWc-J04TKpsFCddLjuF5ZDuBFmokpz4goRgx-FlH-5jCAsFkzkBg
    - type: precision
      value: 0.5385388739946381
      name: Precision Micro
      verified: true
      verifyToken: eyJhbGciOiJFZERTQSIsInR5cCI6IkpXVCJ9.eyJoYXNoIjoiNzRmY2I4YTAzMTRkMjFjNTE1NTEwZDlmZGQ4NDUyYTAxY2JhOTliMDRhNWY3OGY4OWRlNTlkNzcxODc0MDMwYyIsInZlcnNpb24iOjF9.X7ekS-JYOXH5eNmSfKQ_no1rNAbuQ3C0pNYvorPVfcna6RU8n6O6FNQor0AWvatAWdefJG6H3J7_GoC6M5zECw
    - type: precision
      value: 0.5586108016541553
      name: Precision Weighted
      verified: true
      verifyToken: eyJhbGciOiJFZERTQSIsInR5cCI6IkpXVCJ9.eyJoYXNoIjoiMjUwNjMxYjEwMTEzNzAwNzQwZDQwMTRmZDM2ZDk0ZDc3YTUxOTQzNDE5ZWI2NWI4MmJmODAxYTlmN2E0Nzk2MCIsInZlcnNpb24iOjF9.nAO1wRFHMtm5kem9VhuuRg54fpvA2uzwEutjzsnZoyemUHbI2U_1TK_dDmR4bmpPjVnCZt5sF-jEq4oZIaIbDQ
    - type: recall
      value: 0.5385813032215204
      name: Recall Macro
      verified: true
      verifyToken: eyJhbGciOiJFZERTQSIsInR5cCI6IkpXVCJ9.eyJoYXNoIjoiNzVkNjliYTM0Njc3MTUzMDBmYTE5NDRkNzFjNzg2NzA0NzEyMTg4YTlkNGFlZWMxZWUwOGQzYzY1ZGU0ZmIwNyIsInZlcnNpb24iOjF9.cnEbDBJR8m3UqiuzCq_g4RUFLE8BVzXDebKguVrwPgY-Biu4sBFXVQvFyZScsLGEnaHYsE-R8ctTEGDdQONVBw
    - type: recall
      value: 0.5385388739946381
      name: Recall Micro
      verified: true
      verifyToken: eyJhbGciOiJFZERTQSIsInR5cCI6IkpXVCJ9.eyJoYXNoIjoiODZkMmNjZWY4ZDYyYjU3NjQ2ZGNhZjkyNTQyOTg2ZjNmNDgwNDYxYmU2ZDA5M2EwOWRlMjMyYmI4MGU3MGMxNCIsInZlcnNpb24iOjF9.BfMB4_MZ-SYj1YbTES8pqgKNQkNnevSOjAwUqdoL6wsNpsKKWxPHmq0Kt9XufxHoQoyTkGvPfxh-0jEe3B1nBg
    - type: recall
      value: 0.5385388739946381
      name: Recall Weighted
      verified: true
      verifyToken: eyJhbGciOiJFZERTQSIsInR5cCI6IkpXVCJ9.eyJoYXNoIjoiYmE3Yjg3OTVhMjdlMDk1YWFjMWIwNjMyZTA2Yzc3MjBlNjI1YWY5MzE0MjNkMDNiMmU5ZmIxYWExNmViYWE1NSIsInZlcnNpb24iOjF9.S9Bo-wq3wikFS-FqMQerxahu87PJyYx141G5PCWDtOs2wH1nf4texnJYWfHeVCJKZcKmS2RWn5XOjjJ9RoNJAA
    - type: loss
      value: 1.062397837638855
      name: loss
      verified: true
      verifyToken: eyJhbGciOiJFZERTQSIsInR5cCI6IkpXVCJ9.eyJoYXNoIjoiOTFmNDI0ZmQ2YmNlZjJlZTdmZTYwOGVkMjdjMjJkMDIzNzhlOWFiNWQzNjFiMmU5NTdiM2Y1YjYxMjU4ZjQ2ZSIsInZlcnNpb24iOjF9.15RsFRkFpbarlU1L8UyV0o0_5WCveO_mT9CdO0UYwvQsOVjScheJ8fOqHBAC-C-CMTlfFNsmMhNrU_np8c_ZCQ
---

# Indonesian small BigBird model NLI

## Source Code

Source code to create this model and perform benchmark is available at [https://github.com/ilos-vigil/bigbird-small-indonesian](https://github.com/ilos-vigil/bigbird-small-indonesian).

## Model Description

This model is based on [bigbird-small-indonesian](https://huggingface.co/ilos-vigil/bigbird-small-indonesian) and was finetuned on 2 datasets. It is intended to be used for zero-shot text classification.

## How to use

> Inference for ZSC (Zero Shot Classification) task

```py
>>> pipe = pipeline(
...     task='zero-shot-classification',
...     model='./tmp/checkpoint-28832'
... )
>>> pipe(
...     sequences='Fakta nomor 7 akan membuat ada terkejut',
...     candidate_labels=['clickbait', 'bukan clickbait'],
...     hypothesis_template='Judul video ini {}.',
...     multi_label=False
... )
{
 'sequence': 'Fakta nomor 7 akan membuat ada terkejut',
 'labels': ['clickbait', 'bukan clickbait'],
 'scores': [0.6102734804153442, 0.38972654938697815]
}
>>> pipe(
...     sequences='Samsung tuntut balik Apple dengan alasan hak paten teknologi.',
...     candidate_labels=['teknologi', 'olahraga', 'bisnis', 'politik', 'kesehatan', 'kuliner'],
...     hypothesis_template='Kategori berita ini adalah {}.',
...     multi_label=True
... )
{
 'sequence': 'Samsung tuntut balik Apple dengan alasan hak paten teknologi.',
 'labels': ['politik', 'teknologi', 'kesehatan', 'bisnis', 'olahraga', 'kuliner'],
 'scores': [0.7390161752700806, 0.6657379269599915, 0.4459509551525116, 0.38407933712005615, 0.3679264783859253, 0.14181996881961823]
}
```

> Inference for NLI (Natural Language Inference) task

```py
>>> pipe = pipeline(
...     task='text-classification',
...     model='./tmp/checkpoint-28832',
...     return_all_scores=True
... )
>>> pipe({
...     'text': 'Nasi adalah makanan pokok.',  # Premise
...     'text_pair': 'Saya mau makan nasi goreng.'  # Hypothesis
... })
[
 {'label': 'entailment', 'score': 0.25495028495788574},
 {'label': 'neutral', 'score': 0.40920916199684143},
 {'label': 'contradiction', 'score': 0.33584052324295044}
]
>>> pipe({
...     'text': 'Python sering digunakan untuk web development dan AI research.',
...     'text_pair': 'AI research biasanya tidak menggunakan bahasa pemrograman Python.'
... })
[
 {'label': 'entailment', 'score': 0.12508109211921692},
 {'label': 'neutral', 'score': 0.22146646678447723},
 {'label': 'contradiction', 'score': 0.653452455997467}
]
```

## Limitation and bias

This model inherit limitation/bias from it's parent model and 2 datasets used for fine-tuning. And just like most language model, this model is sensitive towards input change. Here's an example.

```py
>>> from transformers import pipeline
>>> pipe = pipeline(
...     task='zero-shot-classification',
...     model='./tmp/checkpoint-28832'
... )
>>> text = 'Resep sate ayam enak dan mudah.'
>>> candidate_labels = ['kuliner', 'olahraga']
>>> pipe(
...     sequences=text,
...     candidate_labels=candidate_labels,
...     hypothesis_template='Kategori judul artikel ini adalah {}.',
...     multi_label=False
... )
{
 'sequence': 'Resep sate ayam enak dan mudah.',
 'labels': ['kuliner', 'olahraga'],
 'scores': [0.7711364030838013, 0.22886358201503754]
}
>>> pipe(
...     sequences=text,
...     candidate_labels=candidate_labels,
...     hypothesis_template='Kelas kalimat ini {}.',
...     multi_label=False
... )
{
 'sequence': 'Resep sate ayam enak dan mudah.',
 'labels': ['kuliner', 'olahraga'],
 'scores': [0.7043636441230774, 0.295636385679245]
}
>>> pipe(
...     sequences=text,
...     candidate_labels=candidate_labels,
...     hypothesis_template='{}.',
...     multi_label=False
... )
{
 'sequence': 'Resep sate ayam enak dan mudah.',
 'labels': ['kuliner', 'olahraga'],
 'scores': [0.5986711382865906, 0.4013288915157318]
}

```

## Training, evaluation and testing data

This model was finetuned with [IndoNLI](https://huggingface.co/datasets/indonli) and [multilingual-NLI-26lang-2mil7](https://huggingface.co/datasets/MoritzLaurer/multilingual-NLI-26lang-2mil7). Although `multilingual-NLI-26lang-2mil7` dataset is machine-translated, this dataset slightly improve result of NLI benchmark and extensively improve result of ZSC benchmark. Both evaluation and testing data is only based on IndoNLI dataset.

## Training Procedure

The model was finetuned on single RTX 3060 with 16 epoch/28832 steps with accumulated batch size 64. AdamW optimizer is used with LR 1e-4, weight decay 0.05, learning rate warmup for first 6% steps (1730 steps) and linear decay of the learning rate afterwards. Take note while model weight on epoch 9 has lowest loss/highest accuracy, it has slightly lower performance on ZSC benchmark. Additional information can be seen on Tensorboard training logs.

## Benchmark as NLI model

Both benchmark show result of 2 different model as additional comparison. Additional benchmark using IndoNLI dataset is available on it's paper [IndoNLI: A Natural Language Inference Dataset for Indonesian](https://aclanthology.org/2021.emnlp-main.821/).

| Model                                      | bigbird-small-indonesian-nli | xlm-roberta-large-xnli | mDeBERTa-v3-base-xnli-multilingual-nli-2mil7 |
| ------------------------------------------ | ---------------------------- | ---------------------- | -------------------------------------------- |
| Parameter                                  | 30.6M                        | 559.9M                 | 278.8M                                       |
| Multilingual                               |                              | V                      | V                                            |
| Finetuned on IndoNLI                       | V                            |                        | V                                            |
| Finetuned on multilingual-NLI-26lang-2mil7 | V                            |                        |                                              |
| Test (Lay)                                 | 0.6888                       | 0.2226                 | 0.8151                                       |
| Test (Expert)                              | 0.5734                       | 0.3505                 | 0.7775                                       |

## Benchmark as ZSC model

[Indonesian-Twitter-Emotion-Dataset](https://github.com/meisaputri21/Indonesian-Twitter-Emotion-Dataset/) is used to perform ZSC benchmark. This benchmark include 4 different parameter which affect performance of each model differently. Hypothesis template for this benchmark is `Kalimat ini mengekspresikan perasaan {}.` and `{}.`. Take note F1 score measurement only calculate label with highest probability.

| Model                                        | Multi-label | Use template | F1 Score     |
| -------------------------------------------- | ----------- | ------------ | ------------ |
| bigbird-small-indonesian-nli                 | V           | V            | 0.3574       |
|                                              | V           |              | 0.3654       |
|                                              |             | V            | 0.3985       |
|                                              |             |              | _0.4160_     |
| xlm-roberta-large-xnli                       | V           | V            | _**0.6292**_ |
|                                              | V           |              | 0.5596       |
|                                              |             | V            | 0.5737       |
|                                              |             |              | 0.5433       |
| mDeBERTa-v3-base-xnli-multilingual-nli-2mil7 | V           | V            | 0.5324       |
|                                              | V           |              | _0.5499_     |
|                                              |             | V            | 0.5269       |
|                                              |             |              | 0.5228       |