File size: 6,835 Bytes
ba2a19f
1db8825
 
ba2a19f
 
 
 
 
 
 
 
 
 
 
 
 
 
1db8825
ba2a19f
 
 
 
 
1db8825
ba2a19f
 
 
 
 
 
 
1db8825
ba2a19f
1db8825
 
ba2a19f
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
---
language:
- en
license: mit
tags:
- generated_from_trainer
datasets:
- glue
metrics:
- accuracy
model-index:
- name: roberta-base-sst2
  results:
  - task:
      name: Text Classification
      type: text-classification
    dataset:
      name: GLUE SST2
      type: glue
      args: sst2
    metrics:
    - name: Accuracy
      type: accuracy
      value: 0.9323394495412844
---

<!-- 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. -->

# roberta-base-sst2

This model is a fine-tuned version of [roberta-base](https://huggingface.co/roberta-base) on the GLUE SST2 dataset.
It achieves the following results on the evaluation set:
- Loss: 0.1952
- Accuracy: 0.9323

## 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: 2e-05
- train_batch_size: 16
- eval_batch_size: 8
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_ratio: 0.06
- num_epochs: 10.0

### Training results

| Training Loss | Epoch | Step  | Validation Loss | Accuracy |
|:-------------:|:-----:|:-----:|:---------------:|:--------:|
| 0.575         | 0.12  | 500   | 0.2665          | 0.9071   |
| 0.2989        | 0.24  | 1000  | 0.2088          | 0.9220   |
| 0.2725        | 0.36  | 1500  | 0.2560          | 0.9243   |
| 0.2814        | 0.48  | 2000  | 0.2016          | 0.9266   |
| 0.2586        | 0.59  | 2500  | 0.2293          | 0.9174   |
| 0.2536        | 0.71  | 3000  | 0.2340          | 0.9323   |
| 0.2494        | 0.83  | 3500  | 0.1952          | 0.9323   |
| 0.2396        | 0.95  | 4000  | 0.2494          | 0.9323   |
| 0.2123        | 1.07  | 4500  | 0.2187          | 0.9381   |
| 0.2042        | 1.19  | 5000  | 0.2812          | 0.9151   |
| 0.2083        | 1.31  | 5500  | 0.2739          | 0.9346   |
| 0.2041        | 1.43  | 6000  | 0.2087          | 0.9381   |
| 0.1969        | 1.54  | 6500  | 0.2590          | 0.9255   |
| 0.1982        | 1.66  | 7000  | 0.2445          | 0.9300   |
| 0.1943        | 1.78  | 7500  | 0.2798          | 0.9266   |
| 0.1848        | 1.9   | 8000  | 0.2844          | 0.9312   |
| 0.1788        | 2.02  | 8500  | 0.2998          | 0.9255   |
| 0.1623        | 2.14  | 9000  | 0.2696          | 0.9392   |
| 0.1499        | 2.26  | 9500  | 0.2533          | 0.9278   |
| 0.1426        | 2.38  | 10000 | 0.2971          | 0.9300   |
| 0.1479        | 2.49  | 10500 | 0.2596          | 0.9358   |
| 0.1405        | 2.61  | 11000 | 0.2945          | 0.9255   |
| 0.1577        | 2.73  | 11500 | 0.4061          | 0.9002   |
| 0.1521        | 2.85  | 12000 | 0.2724          | 0.9335   |
| 0.1426        | 2.97  | 12500 | 0.2712          | 0.9427   |
| 0.1206        | 3.09  | 13000 | 0.2954          | 0.9358   |
| 0.1074        | 3.21  | 13500 | 0.2653          | 0.9392   |
| 0.112         | 3.33  | 14000 | 0.2778          | 0.9346   |
| 0.1147        | 3.44  | 14500 | 0.3705          | 0.9312   |
| 0.1196        | 3.56  | 15000 | 0.2890          | 0.9346   |
| 0.1159        | 3.68  | 15500 | 0.3449          | 0.9266   |
| 0.119         | 3.8   | 16000 | 0.3207          | 0.9335   |
| 0.1268        | 3.92  | 16500 | 0.3235          | 0.9312   |
| 0.1074        | 4.04  | 17000 | 0.3650          | 0.9335   |
| 0.0805        | 4.16  | 17500 | 0.3338          | 0.9381   |
| 0.0838        | 4.28  | 18000 | 0.4302          | 0.9209   |
| 0.0848        | 4.39  | 18500 | 0.4096          | 0.9323   |
| 0.0922        | 4.51  | 19000 | 0.3332          | 0.9369   |
| 0.091         | 4.63  | 19500 | 0.3024          | 0.9438   |
| 0.0977        | 4.75  | 20000 | 0.2674          | 0.9495   |
| 0.0897        | 4.87  | 20500 | 0.3993          | 0.9300   |
| 0.1013        | 4.99  | 21000 | 0.3227          | 0.9289   |
| 0.0671        | 5.11  | 21500 | 0.3374          | 0.9427   |
| 0.0671        | 5.23  | 22000 | 0.4108          | 0.9278   |
| 0.0652        | 5.34  | 22500 | 0.3550          | 0.9381   |
| 0.0664        | 5.46  | 23000 | 0.3398          | 0.9358   |
| 0.0742        | 5.58  | 23500 | 0.3286          | 0.9381   |
| 0.0758        | 5.7   | 24000 | 0.3276          | 0.9312   |
| 0.075         | 5.82  | 24500 | 0.3202          | 0.9369   |
| 0.0686        | 5.94  | 25000 | 0.3481          | 0.9415   |
| 0.0729        | 6.06  | 25500 | 0.3816          | 0.9335   |
| 0.0568        | 6.18  | 26000 | 0.3132          | 0.9381   |
| 0.0529        | 6.29  | 26500 | 0.3757          | 0.9300   |
| 0.0506        | 6.41  | 27000 | 0.3396          | 0.9381   |
| 0.0476        | 6.53  | 27500 | 0.3642          | 0.9404   |
| 0.0555        | 6.65  | 28000 | 0.3430          | 0.9404   |
| 0.0574        | 6.77  | 28500 | 0.3401          | 0.9392   |
| 0.0524        | 6.89  | 29000 | 0.3378          | 0.9346   |
| 0.0492        | 7.01  | 29500 | 0.3833          | 0.9381   |
| 0.039         | 7.13  | 30000 | 0.3347          | 0.9346   |
| 0.0411        | 7.24  | 30500 | 0.4404          | 0.9335   |
| 0.0412        | 7.36  | 31000 | 0.3618          | 0.9381   |
| 0.0477        | 7.48  | 31500 | 0.3806          | 0.9381   |
| 0.0435        | 7.6   | 32000 | 0.3912          | 0.9335   |
| 0.0443        | 7.72  | 32500 | 0.3900          | 0.9392   |
| 0.0421        | 7.84  | 33000 | 0.4152          | 0.9369   |
| 0.0495        | 7.96  | 33500 | 0.3832          | 0.9289   |
| 0.0293        | 8.08  | 34000 | 0.4427          | 0.9346   |
| 0.0253        | 8.19  | 34500 | 0.4425          | 0.9381   |
| 0.0407        | 8.31  | 35000 | 0.4102          | 0.9358   |
| 0.0311        | 8.43  | 35500 | 0.4447          | 0.9369   |
| 0.0291        | 8.55  | 36000 | 0.4612          | 0.9346   |
| 0.035         | 8.67  | 36500 | 0.4241          | 0.9346   |
| 0.0381        | 8.79  | 37000 | 0.4198          | 0.9312   |
| 0.0234        | 8.91  | 37500 | 0.4345          | 0.9369   |
| 0.0311        | 9.03  | 38000 | 0.4558          | 0.9312   |
| 0.028         | 9.14  | 38500 | 0.4245          | 0.9381   |
| 0.0213        | 9.26  | 39000 | 0.4462          | 0.9381   |
| 0.0276        | 9.38  | 39500 | 0.4210          | 0.9381   |
| 0.0183        | 9.5   | 40000 | 0.4310          | 0.9404   |
| 0.0184        | 9.62  | 40500 | 0.4437          | 0.9404   |
| 0.0296        | 9.74  | 41000 | 0.4311          | 0.9392   |
| 0.019         | 9.86  | 41500 | 0.4244          | 0.9415   |
| 0.0245        | 9.98  | 42000 | 0.4270          | 0.9415   |


### Framework versions

- Transformers 4.21.3
- Pytorch 1.7.1
- Datasets 1.18.3
- Tokenizers 0.11.6