exp3_10partition_modelo3000

This model is a fine-tuned version of Helsinki-NLP/opus-mt-es-es on the None dataset. It achieves the following results on the evaluation set:

• Loss: 0.9610
• Bleu Msl: 23.6435
• Bleu 1 Msl: 0.5
• Bleu 2 Msl: 0.0129
• Bleu 3 Msl: 0.0040
• Bleu 4 Msl: 0.0021
• Ter Msl: 25.0
• Bleu Asl: 0
• Bleu 1 Asl: 0
• Bleu 2 Asl: 0
• Bleu 3 Asl: 0
• Bleu 4 Asl: 0
• Ter Asl: 100

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: 0.0001
• train_batch_size: 32
• eval_batch_size: 64
• seed: 42
• optimizer: Use OptimizerNames.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: 30
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Bleu Msl	Bleu 1 Msl	Bleu 2 Msl	Bleu 3 Msl	Bleu 4 Msl	Ter Msl	Bleu Asl	Bleu 1 Asl	Bleu 2 Asl	Bleu 3 Asl	Bleu 4 Asl	Ter Asl
No log	1.0	75	2.4535	0.0	0.0433	0.0038	0.0018	0.0011	66.9477	0	0	0	0	0	100
No log	2.0	150	2.1474	0.0	0.0667	0.0047	0.0021	0.0013	85.1602	0	0	0	0	0	100
No log	3.0	225	2.0004	0.0	0.0633	0.0046	0.0020	0.0012	55.3120	0	0	0	0	0	100
No log	4.0	300	2.0284	0.0	0.07	0.0048	0.0021	0.0013	91.9899	0	0	0	0	0	100
No log	5.0	375	1.9322	0.0	0.0967	0.0057	0.0024	0.0014	54.3002	0	0	0	0	0	100
No log	6.0	450	2.0577	0.0	0.14	0.0068	0.0027	0.0015	48.7352	0	0	0	0	0	100
0.56	7.0	525	2.1728	0.0	0.1033	0.0059	0.0024	0.0014	55.6492	0	0	0	0	0	100
0.56	8.0	600	2.2936	0.0	0.1067	0.0060	0.0024	0.0014	75.2951	0	0	0	0	0	100
0.56	9.0	675	2.0849	77.3055	0.1800	0.0078	0.0029	0.0016	66.6948	0	0	0	0	0	100
0.56	10.0	750	2.0557	0.0	0.1433	0.0069	0.0027	0.0015	56.7454	0	0	0	0	0	100
0.56	11.0	825	2.2431	0.0	0.0933	0.0056	0.0023	0.0014	54.8061	0	0	0	0	0	100
0.56	12.0	900	2.0426	0.0	0.1267	0.0065	0.0026	0.0015	54.3002	0	0	0	0	0	100
0.56	13.0	975	2.1992	0.0	0.1367	0.0068	0.0026	0.0015	141.3153	0	0	0	0	0	100
0.0455	14.0	1050	2.1651	0.0	0.1567	0.0072	0.0028	0.0016	49.1568	0	0	0	0	0	100
0.0455	15.0	1125	2.2092	0.0	0.1233	0.0064	0.0026	0.0015	60.9612	0	0	0	0	0	100
0.0455	16.0	1200	2.1342	0.0	0.1433	0.0069	0.0027	0.0015	55.7336	0	0	0	0	0	100
0.0455	17.0	1275	2.1668	0.0	0.14	0.0068	0.0027	0.0015	56.3238	0	0	0	0	0	100
0.0455	18.0	1350	2.1777	0.0	0.1467	0.0070	0.0027	0.0015	54.7218	0	0	0	0	0	100
0.0455	19.0	1425	2.2567	0.0	0.1533	0.0072	0.0027	0.0016	49.2411	0	0	0	0	0	100
0.0209	20.0	1500	2.2123	0.0	0.1667	0.0075	0.0028	0.0016	48.3137	0	0	0	0	0	100
0.0209	21.0	1575	2.2715	0.0	0.14	0.0068	0.0027	0.0015	51.5177	0	0	0	0	0	100
0.0209	22.0	1650	2.2539	0.0	0.15	0.0071	0.0027	0.0015	51.7707	0	0	0	0	0	100
0.0209	23.0	1725	2.2624	0.0	0.1467	0.0070	0.0027	0.0015	51.6020	0	0	0	0	0	100
0.0209	24.0	1800	2.2808	0.0	0.1467	0.0070	0.0027	0.0015	52.3609	0	0	0	0	0	100
0.0209	25.0	1875	2.2665	0.0	0.15	0.0071	0.0027	0.0015	52.1079	0	0	0	0	0	100
0.0209	26.0	1950	2.2563	0.0	0.14	0.0068	0.0027	0.0015	53.2040	0	0	0	0	0	100
0.0141	27.0	2025	2.2610	0.0	0.1533	0.0072	0.0027	0.0016	51.4334	0	0	0	0	0	100
0.0141	28.0	2100	2.2689	0.0	0.1533	0.0072	0.0027	0.0016	50.8432	0	0	0	0	0	100
0.0141	29.0	2175	2.2707	0.0	0.15	0.0071	0.0027	0.0015	51.1804	0	0	0	0	0	100
0.0141	30.0	2250	2.2715	0.0	0.1533	0.0072	0.0027	0.0016	51.0118	0	0	0	0	0	100

Framework versions

• Transformers 4.49.0
• Pytorch 2.6.0+cu124
• Datasets 3.4.1
• Tokenizers 0.21.1