ehdwns1516 commited on
Commit
52c9221
1 Parent(s): 8901e53

Create README.md

Browse files
Files changed (1) hide show
  1. README.md +37 -0
README.md ADDED
@@ -0,0 +1,37 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ # klue-roberta-base-kornli
2
+
3
+ This model trained with Korean dataset.
4
+ Input premise sentence and hypothesis sentence.
5
+ You can use English, but don't expect accuracy.
6
+ If the context is longer than 1200 characters, the context may be cut in the middle and the result may not come out well.
7
+
8
+ KLUE-RoBERTa-base-KorNLI DEMO: [Ainize DEMO](https://main-klue-roberta-base-kornli-ehdwns1516.endpoint.ainize.ai/)
9
+ KLUE-RoBERTa-base-KorNLI API: [Ainize API](https://ainize.web.app/redirect?git_repo=https://github.com/ehdwns1516/klue-roberta-base_kornli)
10
+
11
+ ## Overview
12
+ Language model: klue/roberta-base
13
+ Language: Korean
14
+ Training data: [kakaobrain KorNLI](https://github.com/kakaobrain/KorNLUDatasets/tree/master/KorNLI)
15
+ Eval data: [kakaobrain KorNLI](https://github.com/kakaobrain/KorNLUDatasets/tree/master/KorNLI)
16
+ Code: See [Ainize Workspace](https://a966119d3186.ngrok.io/notebooks/DJ/KLUE-NLI/klue-roberta-base-kornli.ipynb)
17
+
18
+ ## Usage
19
+ ## In Transformers
20
+
21
+ ```
22
+ from transformers import AutoTokenizer, pipeline
23
+
24
+ tokenizer = AutoTokenizer.from_pretrained("ehdwns1516/klue-roberta-base-kornli")
25
+
26
+ classifier = pipeline(
27
+ "text-classification",
28
+ model="ehdwns1516/klue-roberta-base-kornli",
29
+ return_all_scores=True,
30
+ )
31
+
32
+ premise = "your premise"
33
+ hypothesis = "your hypothesis"
34
+
35
+ result = dict()
36
+ result[0] = classifier(premise + tokenizer.sep_token + hypothesis)[0]
37
+ ```