File size: 5,205 Bytes
64884ff
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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
---
license: mit
tags:
- generated_from_trainer
metrics:
- precision
- recall
- f1
- accuracy
model-index:
- name: pos_final_xlm_en
  results: []
---

<!-- 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. -->

# pos_final_xlm_en

This model is a fine-tuned version of [xlm-roberta-base](https://huggingface.co/xlm-roberta-base) on the None dataset.
It achieves the following results on the evaluation set:
- Loss: 0.0719
- Precision: 0.9686
- Recall: 0.9705
- F1: 0.9695
- Accuracy: 0.9790

## 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: 256
- eval_batch_size: 256
- seed: 42
- gradient_accumulation_steps: 4
- total_train_batch_size: 1024
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_steps: 500
- num_epochs: 40.0
- mixed_precision_training: Native AMP

### Training results

| Training Loss | Epoch | Step | Validation Loss | Precision | Recall | F1     | Accuracy |
|:-------------:|:-----:|:----:|:---------------:|:---------:|:------:|:------:|:--------:|
| No log        | 0.99  | 60   | 3.0062          | 0.2412    | 0.1720 | 0.2008 | 0.3036   |
| No log        | 1.99  | 120  | 0.5353          | 0.8699    | 0.8553 | 0.8625 | 0.8970   |
| No log        | 2.99  | 180  | 0.1312          | 0.9578    | 0.9553 | 0.9566 | 0.9691   |
| No log        | 3.99  | 240  | 0.0981          | 0.9621    | 0.9628 | 0.9625 | 0.9737   |
| No log        | 4.99  | 300  | 0.0853          | 0.9652    | 0.9659 | 0.9655 | 0.9760   |
| No log        | 5.99  | 360  | 0.0788          | 0.9656    | 0.9676 | 0.9666 | 0.9769   |
| No log        | 6.99  | 420  | 0.0745          | 0.9664    | 0.9689 | 0.9677 | 0.9775   |
| No log        | 7.99  | 480  | 0.0718          | 0.9675    | 0.9689 | 0.9682 | 0.9780   |
| 0.7956        | 8.99  | 540  | 0.0707          | 0.9679    | 0.9683 | 0.9681 | 0.9779   |
| 0.7956        | 9.99  | 600  | 0.0686          | 0.9682    | 0.9698 | 0.9690 | 0.9786   |
| 0.7956        | 10.99 | 660  | 0.0686          | 0.9689    | 0.9694 | 0.9692 | 0.9787   |
| 0.7956        | 11.99 | 720  | 0.0680          | 0.9679    | 0.9707 | 0.9693 | 0.9787   |
| 0.7956        | 12.99 | 780  | 0.0685          | 0.9683    | 0.9706 | 0.9694 | 0.9789   |
| 0.7956        | 13.99 | 840  | 0.0695          | 0.9689    | 0.9700 | 0.9694 | 0.9788   |
| 0.7956        | 14.99 | 900  | 0.0703          | 0.9682    | 0.9699 | 0.9690 | 0.9786   |
| 0.7956        | 15.99 | 960  | 0.0719          | 0.9686    | 0.9705 | 0.9695 | 0.9790   |
| 0.051         | 16.99 | 1020 | 0.0735          | 0.9687    | 0.9701 | 0.9694 | 0.9788   |
| 0.051         | 17.99 | 1080 | 0.0747          | 0.9684    | 0.9701 | 0.9692 | 0.9787   |
| 0.051         | 18.99 | 1140 | 0.0761          | 0.9685    | 0.9697 | 0.9691 | 0.9786   |
| 0.051         | 19.99 | 1200 | 0.0774          | 0.9678    | 0.9698 | 0.9688 | 0.9784   |
| 0.051         | 20.99 | 1260 | 0.0796          | 0.9685    | 0.9694 | 0.9690 | 0.9785   |
| 0.051         | 21.99 | 1320 | 0.0796          | 0.9681    | 0.9701 | 0.9691 | 0.9786   |
| 0.051         | 22.99 | 1380 | 0.0820          | 0.9684    | 0.9690 | 0.9687 | 0.9784   |
| 0.051         | 23.99 | 1440 | 0.0829          | 0.9679    | 0.9688 | 0.9683 | 0.9781   |
| 0.0318        | 24.99 | 1500 | 0.0854          | 0.9681    | 0.9690 | 0.9686 | 0.9782   |
| 0.0318        | 25.99 | 1560 | 0.0881          | 0.9677    | 0.9692 | 0.9684 | 0.9782   |
| 0.0318        | 26.99 | 1620 | 0.0893          | 0.9679    | 0.9690 | 0.9685 | 0.9783   |
| 0.0318        | 27.99 | 1680 | 0.0910          | 0.9676    | 0.9691 | 0.9683 | 0.9781   |
| 0.0318        | 28.99 | 1740 | 0.0919          | 0.9684    | 0.9686 | 0.9685 | 0.9783   |
| 0.0318        | 29.99 | 1800 | 0.0933          | 0.9678    | 0.9686 | 0.9682 | 0.9781   |
| 0.0318        | 30.99 | 1860 | 0.0947          | 0.9677    | 0.9688 | 0.9683 | 0.9781   |
| 0.0318        | 31.99 | 1920 | 0.0966          | 0.9678    | 0.9694 | 0.9686 | 0.9783   |
| 0.0318        | 32.99 | 1980 | 0.0974          | 0.9677    | 0.9689 | 0.9683 | 0.9781   |
| 0.0211        | 33.99 | 2040 | 0.0981          | 0.9684    | 0.9693 | 0.9688 | 0.9784   |
| 0.0211        | 34.99 | 2100 | 0.0989          | 0.9681    | 0.9690 | 0.9686 | 0.9783   |
| 0.0211        | 35.99 | 2160 | 0.1008          | 0.9679    | 0.9695 | 0.9687 | 0.9784   |
| 0.0211        | 36.99 | 2220 | 0.1015          | 0.9681    | 0.9689 | 0.9685 | 0.9782   |
| 0.0211        | 37.99 | 2280 | 0.1015          | 0.9677    | 0.9689 | 0.9683 | 0.9781   |
| 0.0211        | 38.99 | 2340 | 0.1024          | 0.9679    | 0.9690 | 0.9684 | 0.9782   |
| 0.0211        | 39.99 | 2400 | 0.1022          | 0.9680    | 0.9690 | 0.9685 | 0.9782   |


### Framework versions

- Transformers 4.25.1
- Pytorch 1.12.0
- Datasets 2.18.0
- Tokenizers 0.13.2