3f3b7e5bbad27cd2bc42cc135cc3179c

This model is a fine-tuned version of Helsinki-NLP/opus-mt-en-sv on the Helsinki-NLP/opus_books [es-nl] dataset. It achieves the following results on the evaluation set:

• Loss: 1.8109
• Data Size: 1.0
• Epoch Runtime: 52.2288
• Bleu: 4.4997

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	7.5534	0	4.3452	0.0658
No log	1	806	6.0208	0.0078	4.7748	0.0691
No log	2	1612	5.2158	0.0156	5.2029	0.1555
No log	3	2418	4.4730	0.0312	6.1287	0.1491
0.1592	4	3224	3.9317	0.0625	7.3686	0.3121
3.7858	5	4030	3.4759	0.125	10.3171	0.7661
3.2801	6	4836	3.0873	0.25	15.3937	1.2030
2.8798	7	5642	2.7229	0.5	26.2236	1.6915
2.5147	8.0	6448	2.3693	1.0	50.5797	2.3233
2.2338	9.0	7254	2.1785	1.0	50.0597	2.8999
2.081	10.0	8060	2.0594	1.0	49.2056	3.2856
1.9754	11.0	8866	1.9799	1.0	51.1090	3.5689
1.8512	12.0	9672	1.9203	1.0	50.9152	3.7557
1.736	13.0	10478	1.8768	1.0	50.1629	3.8585
1.6955	14.0	11284	1.8395	1.0	50.0842	3.9647
1.5994	15.0	12090	1.8146	1.0	49.5349	4.1286
1.5513	16.0	12896	1.8065	1.0	49.3970	4.2160
1.4908	17.0	13702	1.7797	1.0	49.1518	4.3254
1.4014	18.0	14508	1.7789	1.0	48.1806	4.3719
1.4016	19.0	15314	1.7789	1.0	48.3007	4.3991
1.3395	20.0	16120	1.7851	1.0	50.7043	4.4364
1.3102	21.0	16926	1.7747	1.0	51.2518	4.5095
1.2605	22.0	17732	1.7790	1.0	50.9889	4.4983
1.2203	23.0	18538	1.7858	1.0	51.2481	4.5308
1.1826	24.0	19344	1.8005	1.0	51.3522	4.6108
1.1314	25.0	20150	1.8109	1.0	52.2288	4.4997

Framework versions

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