dennispark commited on
Commit
a6cb06f
β€’
1 Parent(s): 96ac545

Update README.md

Browse files
Files changed (1) hide show
  1. README.md +9 -50
README.md CHANGED
@@ -1,50 +1,9 @@
1
- ---
2
- language: ko
3
-
4
- license: cc-by-4.0
5
- ---
6
-
7
- # pko-t5-base
8
- [Source Code](https://github.com/paust-team/pko-t5)
9
-
10
- pko-t5 λŠ” ν•œκ΅­μ–΄ μ „μš© λ°μ΄ν„°λ‘œ ν•™μŠ΅ν•œ [t5 v1.1 λͺ¨λΈ](https://github.com/google-research/text-to-text-transfer-transformer/blob/84f8bcc14b5f2c03de51bd3587609ba8f6bbd1cd/released_checkpoints.md)μž…λ‹ˆλ‹€.
11
-
12
- ν•œκ΅­μ–΄λ₯Ό tokenize ν•˜κΈ° μœ„ν•΄μ„œ sentencepiece λŒ€μ‹  OOV κ°€ μ—†λŠ” BBPE λ₯Ό μ‚¬μš©ν–ˆμœΌλ©° ν•œκ΅­μ–΄ 데이터 (λ‚˜λ¬΄μœ„ν‚€, μœ„ν‚€ν”Όλ””μ•„, λͺ¨λ‘μ˜λ§λ­‰μΉ˜ λ“±..) λ₯Ό T5 의 span corruption task λ₯Ό μ‚¬μš©ν•΄μ„œ unsupervised learning 만 μ μš©ν•˜μ—¬ ν•™μŠ΅μ„ μ§„ν–‰ν–ˆμŠ΅λ‹ˆλ‹€.
13
-
14
- pko-t5 λ₯Ό μ‚¬μš©ν•˜μ‹€ λ•ŒλŠ” λŒ€μƒ task 에 νŒŒμΈνŠœλ‹ν•˜μ—¬ μ‚¬μš©ν•˜μ‹œκΈ° λ°”λžλ‹ˆλ‹€.
15
-
16
- ## Usage
17
- transformers 의 API λ₯Ό μ‚¬μš©ν•˜μ—¬ μ ‘κ·Ό κ°€λŠ₯ν•©λ‹ˆλ‹€. tokenizer λ₯Ό μ‚¬μš©ν• λ•ŒλŠ” `T5Tokenizer` κ°€ μ•„λ‹ˆλΌ `T5TokenizerFast` λ₯Ό μ‚¬μš©ν•΄μ£Όμ‹­μ‹œμ˜€. model 은 T5ForConditionalGeneration λ₯Ό κ·ΈλŒ€λ‘œ ν™œμš©ν•˜μ‹œλ©΄ λ©λ‹ˆλ‹€.
18
-
19
- ### Example
20
- ```python
21
- from transformers import T5TokenizerFast, T5ForConditionalGeneration
22
-
23
- tokenizer = T5TokenizerFast.from_pretrained('paust/pko-t5-base')
24
- model = T5ForConditionalGeneration.from_pretrained('paust/pko-t5-base')
25
-
26
- input_ids = tokenizer(["qa question: λ‹Ήμ‹ μ˜ 이름은 λ¬΄μ—‡μΈκ°€μš”?"]).input_ids
27
- labels = tokenizer(["T5 μž…λ‹ˆλ‹€."]).input_ids
28
- outputs = model(input_ids=input_ids, labels=labels)
29
-
30
- print(f"loss={outputs.loss} logits={outputs.logits}")
31
- ```
32
-
33
-
34
- ## Klue 평가 (dev)
35
-
36
- | | Model | ynat (macro F1) | sts (pearsonr/F1) | nli (acc) | ner (entity-level F1) | re (micro F1) | dp (LAS) | mrc (EM/F1) |
37
- | --- | --- |-----------------| --- | --- | --- | --- | --- | --- |
38
- | | Baseline | **87.30** | **93.20/86.13** | **89.50** | 86.06 | 71.06 | 87.93 | 75.26/- |
39
- | FT | [pko-t5-small](https://huggingface.co/paust/pko-t5-small) (77M) | 86.21 | 77.99/77.01 | 69.20 | 82.60 | 62.95 | 93.15 | 43.81/46.58 |
40
- | FT | [pko-t5-base](https://huggingface.co/paust/pko-t5-base) (250M) | 87.29 | 90.25/83.43 | 79.73 | 87.80 | 72.94 | 97.28 | 61.53/64.74 |
41
- | FT | [pko-t5-large](https://huggingface.co/paust/pko-t5-large) (800M) | 87.12 | 92.05/85.24 | 84.96 | **88.18** | 72.26 | 97.60 | 68.01/71.44 |
42
- | MT | pko-t5-small | 85.85 | 79.12/77.81 | 66.8 | 81.53 | 67.93 | 91.38 | 44.97/48.07 |
43
- | MT | pko-t5-base | 86.86 | 87.61/81.42 | 75.46 | 86.85 | 71.85 | 96.32 | 61.95/65.06 |
44
- | MT | pko-t5-large | 87.25 | 91.05/84.58 | 82.16 | 87.63 | **74.78** | **97.33** | **69.18/71.92** |
45
-
46
- - FT: μ‹±κΈ€νƒœμŠ€ν¬ νŒŒμΈνŠœλ‹ / MT: λ©€ν‹°νƒœμŠ€ν¬ νŒŒμΈνŠœλ‹
47
- - [Baseline](https://arxiv.org/abs/2105.09680): KLUE λ…Όλ¬Έμ—μ„œ μ†Œκ°œλœ dev set 에 λŒ€ν•œ SOTA 점수
48
-
49
- ## License
50
- [PAUST](https://paust.io)μ—μ„œ λ§Œλ“  pko-t5λŠ” [MIT license](https://github.com/paust-team/pko-t5/blob/main/LICENSE) ν•˜μ— κ³΅κ°œλ˜μ–΄ μžˆμŠ΅λ‹ˆλ‹€.
 
1
+ | | Model | ynat (macro F1) | sts (pearsonr/F1) | nli (acc) | ner (entity-level F1) | re (micro F1) | dp (LAS) | mrc (EM/F1) |
2
+ |-----|------------------------------------------------------------------|-----------------|-------------------|-----------|-----------------------|---------------|-----------|-------------|
3
+ | | Baseline | **87.30** | **93.20/86.13** | **89.50** | 86.06 | 71.06 | 87.93 | **75.26/-** |
4
+ | FT | [pko-t5-small](https://huggingface.co/paust/pko-t5-small) (77M) | 86.21 | 77.99/77.01 | 69.20 | 82.60 | 66.46 | 93.15 | 43.81/46.58 |
5
+ | FT | [pko-t5-base](https://huggingface.co/paust/pko-t5-base) (250M) | 87.29 | 90.25/83.43 | 79.73 | 87.80 | 67.23 | 97.28 | 61.53/64.74 |
6
+ | FT | [pko-t5-large](https://huggingface.co/paust/pko-t5-large) (800M) | 87.12 | 92.05/85.24 | 84.96 | **88.18** | **75.17** | **97.60** | 68.01/71.44 |
7
+ | MT | pko-t5-small | 84.54 | 68.50/72/02 | 51.16 | 74.69 | 66.11 | 80.40 | 43.60/46.28 |
8
+ | MT | pko-t5-base | 86.89 | 83.96/80.30 | 72.03 | 85.27 | 66.59 | 95.05 | 61.11/63.94 |
9
+ | MT | pko-t5-large | 87.57 | 91.93/86.29 | 83.63 | 87.41 | 71.34 | 96.99 | 70.70/73.72 |