File size: 3,356 Bytes
d264e50
 
5c3d998
 
d264e50
 
5c3d998
d264e50
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
5b376f0
d264e50
5b376f0
d264e50
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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
---
license: mit
langs:
- multilingual
tags:
- generated_from_trainer
- xnli
datasets:
- xglue
metrics:
- accuracy
model-index:
- name: xlm-v-base-finetuned-xglue-xnli
  results:
  - task:
      name: Text Classification
      type: text-classification
    dataset:
      name: xglue
      type: xglue
      config: xnli
      split: validation.en+validation.ar+validation.bg+validation.de+validation.el+validation.es+validation.fr+validation.hi+validation.ru+validation.sw+validation.th+validation.tr+validation.ur+validation.vi+validation.zh
      args: xnli
    metrics:
    - name: Accuracy
      type: accuracy
      value: 0.7402677376171352
---

<!-- 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. -->

# XLM-V (base) fine-tuned on XNLI

This model is a fine-tuned version of [XLM-V (base)](https://huggingface.co/facebook/xlm-v-base) on the XNLI (XGLUE) dataset.
It achieves the following results on the evaluation set:
- Loss: 0.6511
- Accuracy: 0.7403

## 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: 32
- eval_batch_size: 32
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 2
- mixed_precision_training: Native AMP

### Training results

| Training Loss | Epoch | Step  | Validation Loss | Accuracy |
|:-------------:|:-----:|:-----:|:---------------:|:--------:|
| 1.0994        | 0.08  | 1000  | 1.0966          | 0.3697   |
| 1.0221        | 0.16  | 2000  | 1.0765          | 0.4560   |
| 0.8437        | 0.24  | 3000  | 0.8472          | 0.6179   |
| 0.6997        | 0.33  | 4000  | 0.7650          | 0.6804   |
| 0.6304        | 0.41  | 5000  | 0.7227          | 0.7007   |
| 0.5972        | 0.49  | 6000  | 0.7430          | 0.6977   |
| 0.5886        | 0.57  | 7000  | 0.7365          | 0.7066   |
| 0.5585        | 0.65  | 8000  | 0.6819          | 0.7223   |
| 0.5464        | 0.73  | 9000  | 0.7222          | 0.7046   |
| 0.5289        | 0.81  | 10000 | 0.7290          | 0.7054   |
| 0.5298        | 0.9   | 11000 | 0.6824          | 0.7221   |
| 0.5241        | 0.98  | 12000 | 0.6650          | 0.7268   |
| 0.4806        | 1.06  | 13000 | 0.6861          | 0.7308   |
| 0.4715        | 1.14  | 14000 | 0.6619          | 0.7304   |
| 0.4645        | 1.22  | 15000 | 0.6656          | 0.7284   |
| 0.4443        | 1.3   | 16000 | 0.7026          | 0.7270   |
| 0.4582        | 1.39  | 17000 | 0.7055          | 0.7225   |
| 0.4456        | 1.47  | 18000 | 0.6592          | 0.7361   |
| 0.44          | 1.55  | 19000 | 0.6816          | 0.7329   |
| 0.4419        | 1.63  | 20000 | 0.6772          | 0.7357   |
| 0.4403        | 1.71  | 21000 | 0.6745          | 0.7319   |
| 0.4348        | 1.79  | 22000 | 0.6678          | 0.7338   |
| 0.4355        | 1.87  | 23000 | 0.6614          | 0.7365   |
| 0.4295        | 1.96  | 24000 | 0.6511          | 0.7403   |


### Framework versions

- Transformers 4.26.0
- Pytorch 1.13.1+cu116
- Datasets 2.9.0
- Tokenizers 0.13.2