File size: 2,422 Bytes
455e2c2
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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
---
library_name: transformers
license: mit
base_model: IAmSkyDra/BARTBana_v5
tags:
- generated_from_trainer
metrics:
- sacrebleu
model-index:
- name: BARTBana_Translation_v5
  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. -->

# BARTBana_Translation_v5

This model is a fine-tuned version of [IAmSkyDra/BARTBana_v5](https://huggingface.co/IAmSkyDra/BARTBana_v5) on the None dataset.
It achieves the following results on the evaluation set:
- Loss: 0.5602
- Sacrebleu: 7.0962

## 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: 64
- eval_batch_size: 64
- seed: 42
- optimizer: Use OptimizerNames.ADAMW_TORCH with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
- lr_scheduler_type: linear
- num_epochs: 15
- mixed_precision_training: Native AMP

### Training results

| Training Loss | Epoch | Step  | Validation Loss | Sacrebleu |
|:-------------:|:-----:|:-----:|:---------------:|:---------:|
| 0.8121        | 1.0   | 742   | 0.7206          | 1.9030    |
| 0.7484        | 2.0   | 1484  | 0.6697          | 2.9685    |
| 0.6761        | 3.0   | 2226  | 0.6374          | 4.0287    |
| 0.6557        | 4.0   | 2968  | 0.6162          | 4.8115    |
| 0.6185        | 5.0   | 3710  | 0.6006          | 5.2849    |
| 0.6033        | 6.0   | 4452  | 0.5889          | 5.7585    |
| 0.573         | 7.0   | 5194  | 0.5807          | 6.1391    |
| 0.5573        | 8.0   | 5936  | 0.5758          | 6.3215    |
| 0.5491        | 9.0   | 6678  | 0.5688          | 6.4682    |
| 0.534         | 10.0  | 7420  | 0.5653          | 6.7433    |
| 0.5204        | 11.0  | 8162  | 0.5630          | 6.8459    |
| 0.5144        | 12.0  | 8904  | 0.5620          | 6.9725    |
| 0.5069        | 13.0  | 9646  | 0.5602          | 7.0593    |
| 0.5021        | 14.0  | 10388 | 0.5596          | 7.0546    |
| 0.4929        | 15.0  | 11130 | 0.5602          | 7.0962    |


### Framework versions

- Transformers 4.48.1
- Pytorch 2.5.1+cu124
- Datasets 3.2.0
- Tokenizers 0.21.0