Update README.md
Browse files
README.md
CHANGED
@@ -40,6 +40,89 @@ model-index:
|
|
40 |
- type: f1
|
41 |
value: 86.5
|
42 |
name: F1
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
43 |
---
|
44 |
|
45 |
|
|
|
40 |
- type: f1
|
41 |
value: 86.5
|
42 |
name: F1
|
43 |
+
- task:
|
44 |
+
type: text-similarity
|
45 |
+
name: STS-B
|
46 |
+
dataset:
|
47 |
+
type: glue
|
48 |
+
name: STS-B # Semantic Textual Similarity Benchmark
|
49 |
+
split: stsb
|
50 |
+
metrics:
|
51 |
+
- type: spearmanr
|
52 |
+
value: 83.0
|
53 |
+
name: Spearman Corr
|
54 |
+
- type: pearsonr
|
55 |
+
value: 84.2
|
56 |
+
name: Pearson Corr
|
57 |
+
- task:
|
58 |
+
type: text-classification
|
59 |
+
name: QQP
|
60 |
+
dataset:
|
61 |
+
type: glue
|
62 |
+
name: QQP # Quora Question Pairs
|
63 |
+
split: qqp
|
64 |
+
metrics:
|
65 |
+
- type: f1
|
66 |
+
value: 68.5
|
67 |
+
name: F1
|
68 |
+
- type: accuracy
|
69 |
+
value: 87.7
|
70 |
+
name: Accuracy
|
71 |
+
- task:
|
72 |
+
type: text-classification
|
73 |
+
name: MNLI-m
|
74 |
+
dataset:
|
75 |
+
type: glue
|
76 |
+
name: MNLI-m # MultiNLI Matched
|
77 |
+
split: mnli_matched
|
78 |
+
metrics:
|
79 |
+
- type: accuracy
|
80 |
+
value: 79.9
|
81 |
+
name: Accuracy
|
82 |
+
- task:
|
83 |
+
type: text-classification
|
84 |
+
name: MNLI-mm
|
85 |
+
dataset:
|
86 |
+
type: glue
|
87 |
+
name: MNLI-mm # MultiNLI Matched
|
88 |
+
split: mnli_mismatched
|
89 |
+
metrics:
|
90 |
+
- type: accuracy
|
91 |
+
value: 79.2
|
92 |
+
name: Accuracy
|
93 |
+
- task:
|
94 |
+
type: text-classification
|
95 |
+
name: QNLI
|
96 |
+
dataset:
|
97 |
+
type: glue
|
98 |
+
name: QNLI # Question NLI
|
99 |
+
split: qnli
|
100 |
+
metrics:
|
101 |
+
- type: accuracy
|
102 |
+
value: 89.0
|
103 |
+
name: Accuracy
|
104 |
+
- task:
|
105 |
+
type: text-classification
|
106 |
+
name: RTE
|
107 |
+
dataset:
|
108 |
+
type: glue
|
109 |
+
name: RTE # Recognizing Textual Entailment
|
110 |
+
split: rte
|
111 |
+
metrics:
|
112 |
+
- type: accuracy
|
113 |
+
value: 63.0
|
114 |
+
name: Accuracy
|
115 |
+
- task:
|
116 |
+
type: text-classification
|
117 |
+
name: WNLI
|
118 |
+
dataset:
|
119 |
+
type: glue
|
120 |
+
name: WNLI # Winograd NLI
|
121 |
+
split: wnli
|
122 |
+
metrics:
|
123 |
+
- type: accuracy
|
124 |
+
value: 65.1
|
125 |
+
name: Accuracy
|
126 |
---
|
127 |
|
128 |
|