Add evaluation results on tals/vitaminc dataset

#1
by autoevaluator HF staff - opened
Files changed (1) hide show
  1. README.md +56 -0
README.md CHANGED
@@ -4,6 +4,62 @@ datasets:
4
  - fever
5
  - glue
6
  - tals/vitaminc
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
7
  ---
8
  # Details
9
  Model used in [Get Your Vitamin C! Robust Fact Verification with Contrastive Evidence](https://aclanthology.org/2021.naacl-main.52/) (Schuster et al., NAACL 21`).
 
4
  - fever
5
  - glue
6
  - tals/vitaminc
7
+ model-index:
8
+ - name: tals/albert-base-vitaminc
9
+ results:
10
+ - task:
11
+ type: text-classification
12
+ name: Text Classification
13
+ dataset:
14
+ name: tals/vitaminc
15
+ type: tals/vitaminc
16
+ config: tals--vitaminc
17
+ split: test
18
+ metrics:
19
+ - name: Accuracy
20
+ type: accuracy
21
+ value: 0.4377049477326666
22
+ verified: true
23
+ - name: Precision Macro
24
+ type: precision
25
+ value: 0.3355703134284889
26
+ verified: true
27
+ - name: Precision Micro
28
+ type: precision
29
+ value: 0.4377049477326666
30
+ verified: true
31
+ - name: Precision Weighted
32
+ type: precision
33
+ value: 0.39076019579837573
34
+ verified: true
35
+ - name: Recall Macro
36
+ type: recall
37
+ value: 0.31910363117265955
38
+ verified: true
39
+ - name: Recall Micro
40
+ type: recall
41
+ value: 0.4377049477326666
42
+ verified: true
43
+ - name: Recall Weighted
44
+ type: recall
45
+ value: 0.4377049477326666
46
+ verified: true
47
+ - name: F1 Macro
48
+ type: f1
49
+ value: 0.2979001062785211
50
+ verified: true
51
+ - name: F1 Micro
52
+ type: f1
53
+ value: 0.43770494773266655
54
+ verified: true
55
+ - name: F1 Weighted
56
+ type: f1
57
+ value: 0.39742293746508284
58
+ verified: true
59
+ - name: loss
60
+ type: loss
61
+ value: 1.2437485456466675
62
+ verified: true
63
  ---
64
  # Details
65
  Model used in [Get Your Vitamin C! Robust Fact Verification with Contrastive Evidence](https://aclanthology.org/2021.naacl-main.52/) (Schuster et al., NAACL 21`).