Update README.md
Browse files
README.md
CHANGED
@@ -1,67 +1,31 @@
|
|
1 |
---
|
|
|
2 |
tags:
|
3 |
-
-
|
4 |
-
|
5 |
-
-
|
6 |
-
model-index:
|
7 |
-
- name: kobigbird-finetuned-Encoder-Decoder
|
8 |
-
results: []
|
9 |
-
---
|
10 |
-
|
11 |
-
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
|
12 |
-
should probably proofread and complete it, then remove this comment. -->
|
13 |
-
|
14 |
-
# kobigbird-finetuned-Encoder-Decoder
|
15 |
-
|
16 |
-
This model was trained from scratch on an unknown dataset.
|
17 |
-
It achieves the following results on the evaluation set:
|
18 |
-
- Loss: 4.4011
|
19 |
-
- Rouge1: 12.1436
|
20 |
-
- Rouge2: 2.2747
|
21 |
-
- Rougel: 11.7428
|
22 |
-
- Rougelsum: 11.7408
|
23 |
-
- Gen Len: 20.0
|
24 |
-
|
25 |
-
## Model description
|
26 |
-
|
27 |
-
More information needed
|
28 |
|
29 |
-
|
30 |
-
|
31 |
-
More information needed
|
32 |
-
|
33 |
-
## Training and evaluation data
|
34 |
-
|
35 |
-
More information needed
|
36 |
-
|
37 |
-
## Training procedure
|
38 |
|
39 |
-
|
40 |
|
41 |
-
|
42 |
-
|
43 |
-
- train_batch_size: 1
|
44 |
-
- eval_batch_size: 1
|
45 |
-
- seed: 42
|
46 |
-
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
|
47 |
-
- lr_scheduler_type: linear
|
48 |
-
- num_epochs: 5
|
49 |
-
- mixed_precision_training: Native AMP
|
50 |
|
51 |
-
|
|
|
52 |
|
53 |
-
|
54 |
-
|:-------------:|:-----:|:------:|:---------------:|:-------:|:------:|:-------:|:---------:|:-------:|
|
55 |
-
| 5.8126 | 1.0 | 22194 | 5.7258 | 8.857 | 0.2266 | 8.6768 | 8.703 | 20.0 |
|
56 |
-
| 5.1675 | 2.0 | 44388 | 5.0540 | 7.0446 | 0.5937 | 6.8255 | 6.8637 | 20.0 |
|
57 |
-
| 4.5552 | 3.0 | 66582 | 4.6871 | 10.3238 | 1.1363 | 9.9598 | 9.9394 | 20.0 |
|
58 |
-
| 4.3369 | 4.0 | 88776 | 4.4789 | 11.0189 | 1.6226 | 10.7301 | 10.6951 | 20.0 |
|
59 |
-
| 4.0251 | 5.0 | 110970 | 4.4011 | 12.1436 | 2.2747 | 11.7428 | 11.7408 | 20.0 |
|
60 |
|
|
|
|
|
61 |
|
62 |
-
|
|
|
|
|
|
|
63 |
|
64 |
-
|
65 |
-
|
66 |
-
|
67 |
-
- Tokenizers 0.12.1
|
|
|
1 |
---
|
2 |
+
language: ko
|
3 |
tags:
|
4 |
+
- summarization
|
5 |
+
- bigbird
|
6 |
+
- bart
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
7 |
|
8 |
+
---
|
9 |
+
- This model is a [monologg/kobigbird-bert-base](https://huggingface.co/monologg/kobigbird-bert-base), [ainize/kobart-news](https://huggingface.co/ainize/kobart-news) finetuned on the [daekeun-ml/naver-news-summarization-ko](https://huggingface.co/datasets/daekeun-ml/naver-news-summarization-ko)
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
10 |
|
11 |
+
<<20220917 Commit>>
|
12 |
|
13 |
+
๊ฐ์ธ ์คํฐ๋์ฉ์ผ๋ก ๊ธด ๋ฌธ์ฅ(๋ด์ค ๋ฑ)์ ์์ฝ ๋ชจ๋ธ ํนํ๋ ๋ชจ๋ธ์ ๋ง๋ค๊ธฐ ์ํด BERT๊ธฐ๋ฐ์ KoBigBird ๋ชจ๋ธ์ Encoder Decoder๋ก ๋ณํํ ๋ชจ๋ธ์
๋๋ค.
|
14 |
+
๊ธฐ์กด์ monologg๋์ KoBigBird๋ BERT๊ธฐ๋ฐ์ผ๋ก ๋ฐ์ด๋ ์ฑ๋ฅ์ ์๋ํ์ง๋ง ์์ฑ ์์ฝ ๋ถ๋ถ์ ์์ด์๋ Decoder๊ฐ ์๊ธฐ ๋๋ฌธ์ ์ถ๊ฐ์ ์ผ๋ก Decoder๋ฅผ ๋ถ์์ต๋๋ค.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
15 |
|
16 |
+
๋ง๋ค์๋ ์ด๊ธฐ ๋ชจ๋ธ์ KoBigBird์ Encoder๋ฅผ Decoder๋ก ํ์ฉํ์ฌ ๋ง๋์์ต๋๋ค๋ง, ์์ํ ์ค๋ฅ๋ก ์ธํ์ฌ monologg๋์ KoBigBird-bert-base์ Encoder ๋ถ๋ถ๊ณผ ainize๋์ KoBART-news์ Decoder๋ฅผ ์ด์ด์ ๋ง๋ค์์ต๋๋ค.
|
17 |
+
finetuned ๋ฐ์ดํฐ์
์ผ๋ก daekeun-ml๋์ด ์ ๊ณตํด์ฃผ์ naver-news-summarization-ko ๋ฐ์ดํฐ์
์ ํ์ฉํ์ต๋๋ค.
|
18 |
|
19 |
+
์ดํ AIํ๋ธ์์ ์ ๊ณตํ๋ ์์ฝ ๋ฐ์ดํฐ์
์ผ๋ก ์ถ๊ฐ ํ์ต ์งํ ์์ ์
๋๋ค.
|
|
|
|
|
|
|
|
|
|
|
|
|
20 |
|
21 |
+
์ง์์ ์ผ๋ก ๋ฐ์ ์์ผ ๋์ฑ ์ฌ์ฉํ๊ธฐ ํธํ๊ฒ ๋ง๋ค๊ฒ ์ต๋๋ค.
|
22 |
+
๊ฐ์ฌํฉ๋๋ค.
|
23 |
|
24 |
+
<pre><code>
|
25 |
+
# Python Code
|
26 |
+
from transformers import AutoTokenizer
|
27 |
+
from transformers import AutoModelForSeq2SeqLM
|
28 |
|
29 |
+
tokenizer = AutoTokenizer.from_pretrained("noahkim/KoBigBird-EncoderDecoderModel")
|
30 |
+
model = AutoModelForSeq2SeqLM.from_pretrained("noahkim/KoBigBird-EncoderDecoderModel")
|
31 |
+
</pre></code>
|
|