File size: 1,439 Bytes
23c3f2b
33ec4a3
 
ea676d9
23c3f2b
 
 
 
33ec4a3
 
23c3f2b
 
 
 
 
 
 
 
 
 
ea676d9
23c3f2b
 
 
 
 
 
 
ea676d9
 
23c3f2b
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
ea676d9
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
---
language:
- lfn
license: mit
tags:
- generated_from_trainer
metrics:
- bleu
pipeline_tag: translation
base_model: Helsinki-NLP/opus-mt-mul-en
model-index:
- name: opus-mt-mul-en-finetuned-lfn-to-en
  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. -->

# opus-mt-mul-en-finetuned-lfn-to-en

This model is a fine-tuned version of [Helsinki-NLP/opus-mt-mul-en](https://huggingface.co/Helsinki-NLP/opus-mt-mul-en) on the Tatoeba English-Elefen sentence pair dataset.
It achieves the following results on the evaluation set:
- Loss: 0.6492
- Bleu: 60.0647
- Gen Len: 8.5545

## Model description

Elefen (or Lingua Franca Nova, abbreviated to "LFN") is a simple language designed for international communication.
Its vocabulary is based on Catalan, Spanish, French, Italian and Portuguese. The grammar is very reduced, similar to Romance creoles.

## 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: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 10
- mixed_precision_training: Native AMP

### Framework versions

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