Hugging Face's logo

tunib
/
electra-ko-en-base

Transformers
PyTorch
electra
pretraining
Inference Endpoints
Model card Files Community
1
tunib commited on
Commit
8004b11
•
1 Parent(s): 524c8a1

Update README.md

Browse files
Files changed (1) hide show
  1. README.md +10 -10
README.md CHANGED
@@ -27,26 +27,26 @@ model = AutoModel.from_pretrained('tunib/electra-ko-en-base')
27
  ```
28
 
29
  ## Results on Korean downstream tasks
30
-
31
 
32
  | |**# Params** |**Avg.**| **NSMC**<br/>(acc) | **Naver NER**<br/>(F1) | **PAWS**<br/>(acc) | **KorNLI**<br/>(acc) | **KorSTS**<br/>(spearman) | **Question Pair**<br/>(acc) | **KorQuaD (Dev)**<br/>(EM/F1) |**Korean-Hate-Speech (Dev)**<br/>(F1)|
33
  | :----------------:| :----------------: | :--------------------: | :----------------: | :------------------: | :-----------------------: | :-------------------------: | :---------------------------: | :---------------------------: | :---------------------------: | :----------------: |
34
- |***TUNiB-Electra-ko-base*** | 110M | **85.99** | 90.95 | 87.63 | **84.65** | **82.27** | 85.00 | 95.77 | 64.01 / 90.32 |71.40 |
35
- |***TUNiB-Electra-ko-en-base*** | 133M |84.74 |90.15 | 86.93 | 83.05 | 79.70 | 82.23 | 95.64 | 83.61 / 92.37 |67.86 |
36
  | [KoELECTRA-base-v3](https://github.com/monologg/KoELECTRA) | 110M | 85.92 |90.63 | **88.11** | 84.45 | 82.24 | **85.53** | 95.25 | **84.83 / 93.45** | 67.61 |
37
  | [KcELECTRA-base](https://github.com/Beomi/KcELECTRA) | 124M| 84.75 |**91.71** | 86.90 | 74.80 | 81.65 | 82.65 | **95.78** | 70.60 / 90.11 | **74.49** |
38
- | [KoBERT-base](https://github.com/SKTBrain/KoBERT) | 90M | 81.92 | 89.63 | 86.11 | 80.65 | 79.00 | 79.64 | 93.93 | 52.81 / 80.27 | 66.21 |
39
- | [KcBERT-base](https://github.com/Beomi/KcBERT) | 110M | 79.79 | 89.62 | 84.34 | 66.95 | 74.85 | 75.57 | 93.93 | 60.25 / 84.39 | 68.77 |
40
- | [XLM-Roberta-base](https://github.com/pytorch/fairseq/tree/master/examples/xlmr) | 280M | 83.03 |89.49 | 86.26 | 82.95 | 79.92 | 79.09 | 93.53 | 64.70 / 88.94 | 64.06 |
41
-
42
-
43
 
44
 
45
  ## Results on English downstream tasks
46
 
 
47
 
48
  | |**# Params** | **Avg.** |**CoLA**<br/>(MCC) | **SST**<br/>(Acc) |MRPC<br/>(Acc)| **STS**<br/>(Spearman) | **QQP**<br/>(Acc) | **MNLI**<br/>(Acc) | **QNLI**<br/>(Acc) | **RTE**<br/>(Acc) |
49
  | :----------------:| :----------------: | :--------------------: | :----------------: | :------------------: | :-----------------------: | :-------------------------: | :---------------------------: | :---------------------------: | :---------------------------: | :---------------------------: |
50
- |***TUNiB-Electra-ko-en-base*** | 133M | 85.2| **66.29** | 91.86 | **89.95** | 89.67 | **90.75** | 84.72 | 91.40 |**76.90**|
51
- |[ELECTRA-base](https://github.com/google-research/electra) | 110M | **85.7** | 64.6 | **96.0** | 88.1| **90.2** | 89.5 | **88.5** | **93.1** | 75.2 |
52
  |[BERT-base](https://github.com/google-research/bert) | 110M | 80.8| 52.1 | 93.5 | 84.8| 85.8 | 89.2 | 84.6 | 90.5 | 66.4 |
 
 
27
  ```
28
 
29
  ## Results on Korean downstream tasks
30
+
31
 
32
  | |**# Params** |**Avg.**| **NSMC**<br/>(acc) | **Naver NER**<br/>(F1) | **PAWS**<br/>(acc) | **KorNLI**<br/>(acc) | **KorSTS**<br/>(spearman) | **Question Pair**<br/>(acc) | **KorQuaD (Dev)**<br/>(EM/F1) |**Korean-Hate-Speech (Dev)**<br/>(F1)|
33
  | :----------------:| :----------------: | :--------------------: | :----------------: | :------------------: | :-----------------------: | :-------------------------: | :---------------------------: | :---------------------------: | :---------------------------: | :----------------: |
34
+ |***TUNiB-Electra-ko-base*** | 110M | **85.99** | 90.95 | 87.63 | 84.65 | **82.27** | 85.00 | 95.77 | 64.01 / 90.32 |71.40 |
35
+ |***TUNiB-Electra-ko-en-base*** | 133M |85.34 |90.59 | 87.25 | **84.90** | 80.43 | 83.81 | 94.85 | 83.09 / 92.06 |68.83 |
36
  | [KoELECTRA-base-v3](https://github.com/monologg/KoELECTRA) | 110M | 85.92 |90.63 | **88.11** | 84.45 | 82.24 | **85.53** | 95.25 | **84.83 / 93.45** | 67.61 |
37
  | [KcELECTRA-base](https://github.com/Beomi/KcELECTRA) | 124M| 84.75 |**91.71** | 86.90 | 74.80 | 81.65 | 82.65 | **95.78** | 70.60 / 90.11 | **74.49** |
38
+ | [KoBERT-base](https://github.com/SKTBrain/KoBERT) | 90M | 84.17 | 89.63 | 86.11 | 80.65 | 79.00 | 79.64 | 93.93 | 52.81 / 80.27 | 66.21 |
39
+ | [KcBERT-base](https://github.com/Beomi/KcBERT) | 110M | 81.37 | 89.62 | 84.34 | 66.95 | 74.85 | 75.57 | 93.93 | 60.25 / 84.39 | 68.77 |
40
+ | [XLM-Roberta-base](https://github.com/pytorch/fairseq/tree/master/examples/xlmr) | 280M | 85.74 |89.49 | 86.26 | 82.95 | 79.92 | 79.09 | 93.53 | 64.70 / 88.94 | 64.06 |
 
 
41
 
42
 
43
  ## Results on English downstream tasks
44
 
45
+
46
 
47
  | |**# Params** | **Avg.** |**CoLA**<br/>(MCC) | **SST**<br/>(Acc) |MRPC<br/>(Acc)| **STS**<br/>(Spearman) | **QQP**<br/>(Acc) | **MNLI**<br/>(Acc) | **QNLI**<br/>(Acc) | **RTE**<br/>(Acc) |
48
  | :----------------:| :----------------: | :--------------------: | :----------------: | :------------------: | :-----------------------: | :-------------------------: | :---------------------------: | :---------------------------: | :---------------------------: | :---------------------------: |
49
+ |***TUNiB-Electra-ko-en-base*** | 133M | 85.2| **65.36** | 92.09 | **88.97** | **90.61** | **90.91** | 85.32 | 91.51 |**76.53**|
50
+ |[ELECTRA-base](https://github.com/google-research/electra) | 110M | **85.7** | 64.6 | **96.0** | 88.1| 90.2 | 89.5 | **88.5** | **93.1** | 75.2 |
51
  |[BERT-base](https://github.com/google-research/bert) | 110M | 80.8| 52.1 | 93.5 | 84.8| 85.8 | 89.2 | 84.6 | 90.5 | 66.4 |
52
+