File size: 3,215 Bytes
89e6e1d
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
94512c2
 
 
89e6e1d
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
94512c2
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
---
tags:
- generated_from_trainer
datasets:
- uonlp/CulturaX
metrics:
- accuracy
model-index:
- name: gpt2+ts_cx-en_00000-00009_50k
  results:
  - task:
      name: Causal Language Modeling
      type: text-generation
    dataset:
      name: uonlp/CulturaX en
      type: uonlp/CulturaX
      args: en
    metrics:
    - name: Accuracy
      type: accuracy
      value: 0.3894698710798747
license: mit
language:
- en
---

<!-- 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. -->

# gpt2+ts_cx-en_00000-00009_50k

This model is a fine-tuned version of [](https://huggingface.co/) on the uonlp/CulturaX en dataset.
It achieves the following results on the evaluation set:
- Loss: 3.4121
- Accuracy: 0.3895

## 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: 64
- eval_batch_size: 64
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 1.0

### Training results

| Training Loss | Epoch | Step   | Validation Loss | Accuracy |
|:-------------:|:-----:|:------:|:---------------:|:--------:|
| 4.375         | 0.04  | 10000  | 4.2815          | 0.3111   |
| 4.0754        | 0.08  | 20000  | 3.9984          | 0.3341   |
| 3.9409        | 0.11  | 30000  | 3.8615          | 0.3457   |
| 3.8554        | 0.15  | 40000  | 3.7798          | 0.3531   |
| 3.7973        | 0.19  | 50000  | 3.7210          | 0.3584   |
| 3.7421        | 0.23  | 60000  | 3.6750          | 0.3630   |
| 3.7097        | 0.27  | 70000  | 3.6378          | 0.3664   |
| 3.6741        | 0.3   | 80000  | 3.6061          | 0.3694   |
| 3.6599        | 0.34  | 90000  | 3.5803          | 0.3718   |
| 3.6356        | 0.38  | 100000 | 3.5584          | 0.3741   |
| 3.6131        | 0.42  | 110000 | 3.5423          | 0.3758   |
| 3.5991        | 0.46  | 120000 | 3.5254          | 0.3776   |
| 3.591         | 0.49  | 130000 | 3.5108          | 0.3790   |
| 3.574         | 0.53  | 140000 | 3.4966          | 0.3805   |
| 3.5606        | 0.57  | 150000 | 3.4866          | 0.3815   |
| 3.5516        | 0.61  | 160000 | 3.4739          | 0.3828   |
| 3.5423        | 0.64  | 170000 | 3.4650          | 0.3838   |
| 3.5298        | 0.68  | 180000 | 3.4560          | 0.3847   |
| 3.5287        | 0.72  | 190000 | 3.4479          | 0.3857   |
| 3.5187        | 0.76  | 200000 | 3.4408          | 0.3863   |
| 3.5157        | 0.8   | 210000 | 3.4339          | 0.3870   |
| 3.5042        | 0.83  | 220000 | 3.4286          | 0.3876   |
| 3.5033        | 0.87  | 230000 | 3.4229          | 0.3883   |
| 3.501         | 0.91  | 240000 | 3.4188          | 0.3888   |
| 3.4946        | 0.95  | 250000 | 3.4149          | 0.3892   |
| 3.4971        | 0.99  | 260000 | 3.4126          | 0.3894   |


### Framework versions

- Transformers 4.37.1
- Pytorch 2.1.2+cu121
- Datasets 2.16.1
- Tokenizers 0.15.1