JesseStover
commited on
Commit
β’
8f83403
1
Parent(s):
e5af8b5
Create README.md
Browse files
README.md
ADDED
@@ -0,0 +1,40 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
---
|
2 |
+
language:
|
3 |
+
- ko
|
4 |
+
---
|
5 |
+
|
6 |
+
The L2AI-dictionary model is fine-tuned for multiple choice, specifically for selecting the best dictionary definition of a given word in a sentence. Below is an example usage:
|
7 |
+
|
8 |
+
```python
|
9 |
+
import numpy as np
|
10 |
+
import torch
|
11 |
+
from transformers import AutoModelForMultipleChoice, AutoTokenizer
|
12 |
+
|
13 |
+
model_name = "JesseStover/L2AI-dictionary-klue-bert-base"
|
14 |
+
tokenizer = AutoTokenizer.from_pretrained(model_name)
|
15 |
+
model = AutoModelForMultipleChoice.from_pretrained(model_name)
|
16 |
+
model.to(torch.device("cuda" if torch.cuda.is_available() else "cpu"))
|
17 |
+
|
18 |
+
prompts = "\"κ°μμ§λ λ½μ‘λ½μ‘νλ€.\"μ μλ \"κ°μμ§\"μ μ μλ "
|
19 |
+
candidates = [
|
20 |
+
"\"(λͺ
μ¬) κ°μ μλΌ\"μμ.",
|
21 |
+
"\"(λͺ
μ¬) λΆλͺ¨λ ν μλ²μ§, ν λ¨Έλκ° μμμ΄λ μμ£Όλ₯Ό κ·μ¬μνλ©΄μ λΆλ₯΄λ λ§\"μ΄μμ."
|
22 |
+
]
|
23 |
+
|
24 |
+
inputs = tokenizer(
|
25 |
+
[[prompt, candidate] for candidate in candidates],
|
26 |
+
return_tensors="pt",
|
27 |
+
padding=True
|
28 |
+
)
|
29 |
+
|
30 |
+
labels = torch.tensor(0).unsqueeze(0)
|
31 |
+
|
32 |
+
with torch.no_grad():
|
33 |
+
outputs = model(
|
34 |
+
**{k: v.unsqueeze(0) for k, v in inputs.items()}, labels=labels
|
35 |
+
)
|
36 |
+
|
37 |
+
print({i: float(x) for i, x in enumerate(outputs.logits.softmax(1)[0])})
|
38 |
+
```
|
39 |
+
|
40 |
+
Training data was procured under Creative Commons [CC BY-SA 2.0 KR DEED](https://creativecommons.org/licenses/by-sa/2.0/kr/) from the National Institute of Korean Language's [Basic Korean Dictionary](https://krdict.korean.go.kr) and [Standard Korean Dictionary](https://stdict.korean.go.kr/).
|