File size: 1,599 Bytes
89cbf44
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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
---
license: mit
base_model: yomilimi/Jeolla_encoder
tags:
- generated_from_trainer
metrics:
- bleu
model-index:
- name: Je_gyeongsang_test
  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. -->

# Je_gyeongsang_test

This model is a fine-tuned version of [yomilimi/Jeolla_encoder](https://huggingface.co./yomilimi/Jeolla_encoder) on the None dataset.
It achieves the following results on the evaluation set:
- Loss: 0.0061
- Bleu: 91.3211
- Gen Len: 13.3556

## 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: 3
- mixed_precision_training: Native AMP

### Training results

| Training Loss | Epoch | Step  | Validation Loss | Bleu    | Gen Len |
|:-------------:|:-----:|:-----:|:---------------:|:-------:|:-------:|
| 0.008         | 1.0   | 12910 | 0.0073          | 91.096  | 13.3545 |
| 0.0062        | 2.0   | 25820 | 0.0064          | 91.3089 | 13.3544 |
| 0.0055        | 3.0   | 38730 | 0.0061          | 91.3211 | 13.3556 |


### Framework versions

- Transformers 4.42.0.dev0
- Pytorch 2.2.1+cu121
- Datasets 2.19.2
- Tokenizers 0.19.1