translation_model

This model was trained from scratch on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 1.8172
• Bleu: 1.4433
• Gen Len: 18.005

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: 16
• eval_batch_size: 16
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 2
• mixed_precision_training: Native AMP

Training results

Epoch	Step	Training Loss	Validation Loss	Bleu	Gen Len
1.0	4674	2.5538	2.2471	0.7389	18.1428
2.0	9348	2.4606	2.1615	0.8787	18.1105
3.0	14022	2.3877	2.0981	0.9630	18.0917
4.0	18696	2.3433	2.0538	1.0353	18.0855
5.0	23370	2.2914	2.0186	1.0902	18.0692
6.0	28044	2.2565	1.9939	1.1566	18.0789
7.0	32718	2.2539	1.9756	1.1935	18.0686
8.0	37392	2.2262	1.9628	1.1998	18.0759
9.0	42066	2.2097	1.9556	1.2209	18.0706
10.0	46740	2.2021	1.9529	1.2244	18.0659
11.0	51414	2.1283	1.8698	1.3343	18.024
12.0	56088	2.1211	1.8516	1.3656	18.0194
13.0	60762	2.1063	1.8461	1.3828	18.0112
14.0	65436	2.0545	1.8267	1.4183	18.0116
15.0	70110	2.0645	1.8172	1.4433	18.005

Framework versions

• Transformers 4.40.1
• Pytorch 2.2.1+cu121
• Datasets 2.19.1
• Tokenizers 0.19.1