File size: 4,596 Bytes
2f97ee7
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
---
license: apache-2.0
tags:
- generated_from_trainer
datasets:
- stereoset
metrics:
- accuracy
model-index:
- name: bert-base-uncased_stereoset_finetuned
  results:
  - task:
      name: Text Classification
      type: text-classification
    dataset:
      name: stereoset
      type: stereoset
      config: intersentence
      split: validation
      args: intersentence
    metrics:
    - name: Accuracy
      type: accuracy
      value: 0.7260596546310832
---

<!-- This model card has been generated automatically according to the information the Trainer had access to. You
should probably proofread and complete it, then remove this comment. -->

# bert-base-uncased_stereoset_finetuned

This model is a fine-tuned version of [bert-base-uncased](https://huggingface.co/bert-base-uncased) on the stereoset dataset.
It achieves the following results on the evaluation set:
- Loss: 1.3464
- Accuracy: 0.7261

## Model description

More information needed

## Intended uses & limitations

More information needed

## Training and evaluation data

More information needed

## Training procedure

### Training hyperparameters

The following hyperparameters were used during training:
- learning_rate: 5e-05
- train_batch_size: 128
- eval_batch_size: 64
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 10

### Training results

| Training Loss | Epoch | Step | Validation Loss | Accuracy |
|:-------------:|:-----:|:----:|:---------------:|:--------:|
| No log        | 0.21  | 5    | 0.6832          | 0.5565   |
| No log        | 0.42  | 10   | 0.6945          | 0.4741   |
| No log        | 0.62  | 15   | 0.6659          | 0.6224   |
| No log        | 0.83  | 20   | 0.6337          | 0.6758   |
| No log        | 1.04  | 25   | 0.6019          | 0.6695   |
| No log        | 1.25  | 30   | 0.5797          | 0.7096   |
| No log        | 1.46  | 35   | 0.5562          | 0.7166   |
| No log        | 1.67  | 40   | 0.5497          | 0.7363   |
| No log        | 1.88  | 45   | 0.5382          | 0.7418   |
| No log        | 2.08  | 50   | 0.5356          | 0.7418   |
| No log        | 2.29  | 55   | 0.5690          | 0.7316   |
| No log        | 2.5   | 60   | 0.5778          | 0.7418   |
| No log        | 2.71  | 65   | 0.5695          | 0.7386   |
| No log        | 2.92  | 70   | 0.5765          | 0.7386   |
| No log        | 3.12  | 75   | 0.6079          | 0.7363   |
| No log        | 3.33  | 80   | 0.6919          | 0.7370   |
| No log        | 3.54  | 85   | 0.7396          | 0.7339   |
| No log        | 3.75  | 90   | 0.7109          | 0.7339   |
| No log        | 3.96  | 95   | 0.7246          | 0.7308   |
| No log        | 4.17  | 100  | 0.7502          | 0.7292   |
| No log        | 4.38  | 105  | 0.8222          | 0.7331   |
| No log        | 4.58  | 110  | 0.8508          | 0.7268   |
| No log        | 4.79  | 115  | 0.8995          | 0.7378   |
| No log        | 5.0   | 120  | 0.8797          | 0.7323   |
| No log        | 5.21  | 125  | 0.9254          | 0.7370   |
| No log        | 5.42  | 130  | 0.9863          | 0.7292   |
| No log        | 5.62  | 135  | 1.0044          | 0.7198   |
| No log        | 5.83  | 140  | 1.0236          | 0.7339   |
| No log        | 6.04  | 145  | 1.0705          | 0.7355   |
| No log        | 6.25  | 150  | 1.0734          | 0.7323   |
| No log        | 6.46  | 155  | 1.1066          | 0.7300   |
| No log        | 6.67  | 160  | 1.1166          | 0.7292   |
| No log        | 6.88  | 165  | 1.1258          | 0.7370   |
| No log        | 7.08  | 170  | 1.1972          | 0.7300   |
| No log        | 7.29  | 175  | 1.1682          | 0.7268   |
| No log        | 7.5   | 180  | 1.2221          | 0.7166   |
| No log        | 7.71  | 185  | 1.2813          | 0.7151   |
| No log        | 7.92  | 190  | 1.3180          | 0.7214   |
| No log        | 8.12  | 195  | 1.3224          | 0.7166   |
| No log        | 8.33  | 200  | 1.3252          | 0.7135   |
| No log        | 8.54  | 205  | 1.3205          | 0.7221   |
| No log        | 8.75  | 210  | 1.3266          | 0.7245   |
| No log        | 8.96  | 215  | 1.3318          | 0.7206   |
| No log        | 9.17  | 220  | 1.3345          | 0.7237   |
| No log        | 9.38  | 225  | 1.3378          | 0.7245   |
| No log        | 9.58  | 230  | 1.3422          | 0.7261   |
| No log        | 9.79  | 235  | 1.3453          | 0.7261   |
| No log        | 10.0  | 240  | 1.3464          | 0.7261   |


### Framework versions

- Transformers 4.26.1
- Pytorch 1.13.1
- Datasets 2.9.0
- Tokenizers 0.13.2