snumin44 commited on
Commit
b225c11
ยท
verified ยท
1 Parent(s): b4c527a

Update README.md

Browse files
Files changed (1) hide show
  1. README.md +27 -2
README.md CHANGED
@@ -10,13 +10,38 @@ metrics:
10
 
11
  # ๐ŸŠ DPR-KO
12
 
 
 
13
  **ํ•œ๊ตญ์–ด DPR ๋ชจ๋ธ (Question Encoder)** ์ž…๋‹ˆ๋‹ค.
14
  Facebook์˜ DPR ์ฝ”๋“œ์™€๋Š” ์ „ํ˜€ ๋‹ค๋ฅธ ์ƒˆ๋กœ์šด ์ฝ”๋“œ๋กœ ํ•™์Šต๋˜์—ˆ์Šต๋‹ˆ๋‹ค.
15
  Dense Vector ๊ธฐ๋ฐ˜์˜ Semantic Search์— ์‚ฌ์šฉํ•  ์ˆ˜ ์žˆ์Šต๋‹ˆ๋‹ค.
16
  ์งˆ๋ฌธ์€ Question Encoder๋กœ, ํ…์ŠคํŠธ๋Š” Context Encoder๋ฅผ ์ด์šฉํ•ด ์ธ์ฝ”๋”ฉํ•ฉ๋‹ˆ๋‹ค.
17
 
18
  - Github: [https://github.com/snumin44/DPR-KO](https://github.com/snumin44/DPR-KO)
19
- - Context Encoder:
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
20
 
21
-
 
22
 
 
10
 
11
  # ๐ŸŠ DPR-KO
12
 
13
+ ## 1. Intro
14
+
15
  **ํ•œ๊ตญ์–ด DPR ๋ชจ๋ธ (Question Encoder)** ์ž…๋‹ˆ๋‹ค.
16
  Facebook์˜ DPR ์ฝ”๋“œ์™€๋Š” ์ „ํ˜€ ๋‹ค๋ฅธ ์ƒˆ๋กœ์šด ์ฝ”๋“œ๋กœ ํ•™์Šต๋˜์—ˆ์Šต๋‹ˆ๋‹ค.
17
  Dense Vector ๊ธฐ๋ฐ˜์˜ Semantic Search์— ์‚ฌ์šฉํ•  ์ˆ˜ ์žˆ์Šต๋‹ˆ๋‹ค.
18
  ์งˆ๋ฌธ์€ Question Encoder๋กœ, ํ…์ŠคํŠธ๋Š” Context Encoder๋ฅผ ์ด์šฉํ•ด ์ธ์ฝ”๋”ฉํ•ฉ๋‹ˆ๋‹ค.
19
 
20
  - Github: [https://github.com/snumin44/DPR-KO](https://github.com/snumin44/DPR-KO)
21
+ - Context Encoder: [https://huggingface.co/snumin44/biencoder-ko-bert-context](https://huggingface.co/snumin44/biencoder-ko-bert-context)
22
+
23
+
24
+ ## 2. Experiment settings
25
+
26
+ - ๋ฒ ์ด์Šค ๋ชจ๋ธ: klue/bert-base
27
+ - ๋ฐ์ดํ„ฐ ์…‹: KorQuad v1
28
+ - ์œ„ํ‚ค ๋คํ”„: kowiki-latest-pages-articles.xml.bz2 (2024/07/23)
29
+ - ์ฒญํฌ ๋‹น ๋ฌธ์žฅ: 5
30
+ - ์ „์ฒด ์ฒญํฌ: ์•ฝ 160 ๋งŒ
31
+ - BM25 ๊ฐ€์ค‘์น˜: 0.3
32
+ - 1 A100 GPU
33
+
34
+ ## 3. Performance
35
+
36
+ |(%)|BM25 (w/o DPR-KO)|DPR-KO (w/o BM25)|DPR-KO (with BM25)|
37
+ |:---:|:---:|:---:|:---:|
38
+ |Top1 Acc|36.25 |**48.98** |71.16 |
39
+ |Top5 Acc|51.61 |**71.16** |86.75 |
40
+ |Top10 Acc|57.34 |**77.05** |90.28 |
41
+ |Top20 Acc|62.40 |**82.09** |92.66 |
42
+ |Top50 Acc|68.46 |**87.03** |94.86 |
43
+ |Top100 Acc|72.48 |**90.23** |96.02 |
44
 
45
+ โ€ป BM25๋ชจ๋ธ์€ ํ•œ๊ตญ์–ด ์œ„ํ‚คํ”ผ๋””์•„ ์ „์ฒด ํ…์ŠคํŠธ๋กœ ํ•™์Šตํ•œ ๋ชจ๋ธ์ž…๋‹ˆ๋‹ค.
46
+ โ€ป ์ž์„ธํ•œ ์ฝ”๋“œ๋Š” Github ๋ฅผ ์ฐธ๊ณ ํ•ด์ฃผ์„ธ์š”.
47