File size: 1,839 Bytes
cbc4ff8
2ed6a83
 
 
cbc4ff8
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
2ed6a83
cbc4ff8
ceddb6b
 
 
 
2ed6a83
ceddb6b
cbc4ff8
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
2ed6a83
 
 
cbc4ff8
b625cb9
cbc4ff8
 
b625cb9
cbc4ff8
 
 
2ed6a83
 
ceddb6b
 
 
cbc4ff8
 
 
 
2ed6a83
b625cb9
cbc4ff8
b625cb9
2ed6a83
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
---
library_name: peft
license: mit
base_model: facebook/m2m100_418M
tags:
- generated_from_trainer
metrics:
- bleu
- f1
- wer
model-index:
- name: m2m_trial1
  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. -->

# m2m_trial1

This model is a fine-tuned version of [facebook/m2m100_418M](https://huggingface.co/facebook/m2m100_418M) on the None dataset.
It achieves the following results on the evaluation set:
- Bleu: 0.8239
- F1: 0.9229
- Wer: 0.0824
- Cer: 0.0262
- Meteor: 0.9148
- Loss: 6.1042

## 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: 0.0005
- train_batch_size: 32
- eval_batch_size: 32
- seed: 42
- optimizer: Use 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: 3
- mixed_precision_training: Native AMP

### Training results

| Training Loss | Epoch | Step  | Bleu   | F1     | Wer    | Cer    | Meteor | Validation Loss |
|:-------------:|:-----:|:-----:|:------:|:------:|:------:|:------:|:------:|:---------------:|
| 6.1256        | 1.0   | 12500 | 0.7992 | 0.9121 | 0.0950 | 0.0308 | 0.9022 | 6.1147          |
| 6.1187        | 2.0   | 25000 | 0.8172 | 0.9198 | 0.0868 | 0.0281 | 0.9112 | 6.1067          |
| 6.0999        | 3.0   | 37500 | 0.8239 | 0.9229 | 0.0824 | 0.0262 | 0.9148 | 6.1042          |


### Framework versions

- PEFT 0.15.2
- Transformers 4.50.0
- Pytorch 2.6.0+cu124
- Datasets 3.4.1
- Tokenizers 0.21.1