File size: 1,432 Bytes
9799a8a
 
9f07e15
9799a8a
ab39236
 
9799a8a
 
 
 
ab39236
 
 
 
 
 
 
 
e6f3cda
ab39236
 
 
 
e6f3cda
9799a8a
 
 
 
 
 
 
ab39236
9799a8a
e6f3cda
 
9799a8a
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
e6f3cda
9799a8a
ab39236
9799a8a
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
---
tags:
- translation3
- generated_from_trainer
datasets:
- ccmatrix
metrics:
- bleu
model-index:
- name: opus-mt-tc-big-ar-en
  results:
  - task:
      name: Sequence-to-sequence Language Modeling
      type: text2text-generation
    dataset:
      name: ccmatrix
      type: ccmatrix
      config: ar-en
      split: train[498900:700000]
      args: ar-en
    metrics:
    - name: Bleu
      type: bleu
      value: 58.426991354513206
---

<!-- 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-tc-big-ar-en

This model was trained from scratch on the ccmatrix dataset.
It achieves the following results on the evaluation set:
- Loss: 0.6490
- Bleu: 58.4270

## 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: 64
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 3
- mixed_precision_training: Native AMP

### Training results



### Framework versions

- Transformers 4.27.3
- Pytorch 1.13.1+cu116
- Datasets 2.10.1
- Tokenizers 0.13.2