a4d1f3d879817e26a5dfd54a4cac80b3

This model is a fine-tuned version of Helsinki-NLP/opus-mt-en-ru on the Helsinki-NLP/opus_books [fi-no] dataset. It achieves the following results on the evaluation set:

• Loss: 2.8062
• Data Size: 1.0
• Epoch Runtime: 6.7818
• Bleu: 1.4131

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: 5e-05
• train_batch_size: 8
• eval_batch_size: 8
• seed: 42
• distributed_type: multi-GPU
• num_devices: 4
• total_train_batch_size: 32
• total_eval_batch_size: 32
• optimizer: Use adamw_torch with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: constant
• num_epochs: 50

Training results

Training Loss	Epoch	Step	Validation Loss	Data Size	Epoch Runtime	Bleu
No log	0	0	8.1356	0	1.0873	0.0127
No log	1	85	7.3057	0.0078	1.8890	0.0110
No log	2	170	6.6622	0.0156	1.4556	0.0172
No log	3	255	6.1660	0.0312	1.4796	0.0206
No log	4	340	5.5373	0.0625	1.8547	0.0369
0.3757	5	425	4.9089	0.125	2.1591	0.0581
0.3757	6	510	4.2802	0.25	3.0434	0.1016
1.3649	7	595	3.7764	0.5	4.0210	0.2707
3.717	8.0	680	3.4107	1.0	6.6854	0.5826
3.314	9.0	765	3.2178	1.0	6.2309	0.7466
3.1192	10.0	850	3.0836	1.0	6.2849	0.7993
2.9935	11.0	935	2.9972	1.0	6.1519	0.9247
2.8241	12.0	1020	2.9433	1.0	6.7706	1.0162
2.7129	13.0	1105	2.8877	1.0	6.2204	1.0726
2.6121	14.0	1190	2.8476	1.0	6.4070	1.0808
2.5191	15.0	1275	2.8171	1.0	6.1035	1.1208
2.4176	16.0	1360	2.7938	1.0	6.2285	1.2016
2.3275	17.0	1445	2.7758	1.0	6.5607	1.2120
2.2484	18.0	1530	2.7729	1.0	6.0973	1.2572
2.1585	19.0	1615	2.7696	1.0	6.1320	1.2579
2.0738	20.0	1700	2.7718	1.0	6.4671	1.3929
1.9766	21.0	1785	2.7753	1.0	6.2386	1.3285
1.9247	22.0	1870	2.7877	1.0	6.5312	1.3564
1.834	23.0	1955	2.8062	1.0	6.7818	1.4131

Framework versions

• Transformers 4.57.0
• Pytorch 2.8.0+cu128
• Datasets 4.2.0
• Tokenizers 0.22.1