File size: 3,072 Bytes
d009825
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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
---
tags:
- generated_from_trainer
datasets:
- ccmatrix
metrics:
- bleu
model-index:
- name: opus-mt-id-en-ccmatrix-no-warmup
  results:
  - task:
      name: Sequence-to-sequence Language Modeling
      type: text2text-generation
    dataset:
      name: ccmatrix
      type: ccmatrix
      config: en-id
      split: train
      args: en-id
    metrics:
    - name: Bleu
      type: bleu
      value: 56.6595
---

<!-- 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. -->

# opus-mt-id-en-ccmatrix-no-warmup

This model was trained from scratch on the ccmatrix dataset.
It achieves the following results on the evaluation set:
- Loss: 0.9795
- Bleu: 56.6595

## 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: 0.0001
- 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: 25

### Training results

| Training Loss | Epoch | Step   | Validation Loss | Bleu    |
|:-------------:|:-----:|:------:|:---------------:|:-------:|
| 0.8307        | 1.0   | 28125  | 0.8213          | 50.8248 |
| 0.7031        | 2.0   | 56250  | 0.7851          | 52.0712 |
| 0.6426        | 3.0   | 84375  | 0.7716          | 52.6974 |
| 0.5976        | 4.0   | 112500 | 0.7645          | 53.0742 |
| 0.5599        | 5.0   | 140625 | 0.7580          | 53.6526 |
| 0.5265        | 6.0   | 168750 | 0.7610          | 53.9336 |
| 0.4962        | 7.0   | 196875 | 0.7592          | 54.3884 |
| 0.4677        | 8.0   | 225000 | 0.7611          | 54.4904 |
| 0.4404        | 9.0   | 253125 | 0.7680          | 54.6679 |
| 0.4148        | 10.0  | 281250 | 0.7760          | 54.8934 |
| 0.3901        | 11.0  | 309375 | 0.7853          | 55.1644 |
| 0.3662        | 12.0  | 337500 | 0.7945          | 55.3496 |
| 0.3432        | 13.0  | 365625 | 0.8071          | 55.4056 |
| 0.3211        | 14.0  | 393750 | 0.8149          | 55.5914 |
| 0.2995        | 15.0  | 421875 | 0.8314          | 55.8594 |
| 0.2789        | 16.0  | 450000 | 0.8504          | 55.5482 |
| 0.2591        | 17.0  | 478125 | 0.8637          | 55.8965 |
| 0.24          | 18.0  | 506250 | 0.8789          | 55.9817 |
| 0.222         | 19.0  | 534375 | 0.8956          | 56.046  |
| 0.2047        | 20.0  | 562500 | 0.9096          | 56.1539 |
| 0.1888        | 21.0  | 590625 | 0.9279          | 56.2939 |
| 0.1739        | 22.0  | 618750 | 0.9409          | 56.4546 |
| 0.1606        | 23.0  | 646875 | 0.9580          | 56.5525 |
| 0.1488        | 24.0  | 675000 | 0.9713          | 56.6069 |
| 0.1392        | 25.0  | 703125 | 0.9795          | 56.6595 |


### Framework versions

- Transformers 4.26.1
- Pytorch 2.0.0
- Datasets 2.10.1
- Tokenizers 0.11.0