exp5_10partition_modelo12000

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: 0.5181
• Bleu Msl: 100.0000
• Bleu 1 Msl: 0.2933
• Bleu 2 Msl: 0.0099
• Bleu 3 Msl: 0.0034
• Bleu 4 Msl: 0.0018
• Ter Msl: 62.3904
• Bleu Asl: 100.0000
• Bleu 1 Asl: 0.5289
• Bleu 2 Asl: 0.0077
• Bleu 3 Asl: 0.0020
• Bleu 4 Asl: 0.0009
• Ter Asl: 9.6314

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	300	0.4552	84.0896	0.0867	0.0054	0.0023	0.0013	717.8919	0.0	0.5744	0.0080	0.0020	0.0009	7.3645
0.4443	2.0	600	0.3655	100.0000	0.1033	0.0059	0.0024	0.0014	57.2188	0.0	0.5856	0.0081	0.0021	0.0009	6.3300
0.4443	3.0	900	0.3487	59.4604	0.1333	0.0067	0.0026	0.0015	78.8308	0.0	0.5989	0.0082	0.0021	0.0010	6.6010
0.0984	4.0	1200	0.3369	100.0000	0.1600	0.0073	0.0028	0.0016	48.8043	0.0	0.5856	0.0081	0.0021	0.0009	7.0936
0.0592	5.0	1500	0.3564	23.3569	0.1433	0.0069	0.0027	0.0015	54.2073	0.0	0.5133	0.0076	0.0020	0.0009	10.7882
0.0592	6.0	1800	0.3252	100.0000	0.2033	0.0082	0.0030	0.0017	50.3986	0.0	0.5233	0.0076	0.0020	0.0009	9.8768
0.0357	7.0	2100	0.3320	70.7107	0.1733	0.0076	0.0029	0.0016	51.7272	0.0	0.5533	0.0078	0.0020	0.0009	8.1034
0.0357	8.0	2400	0.3809	50.0000	0.1200	0.0063	0.0025	0.0015	56.5102	0.0	0.5544	0.0079	0.0020	0.0009	8.9655
0.0268	9.0	2700	0.2994	28.9178	0.2167	0.0085	0.0031	0.0017	53.2329	0.0	0.5189	0.0076	0.0020	0.0009	10.4680
0.02	10.0	3000	0.3532	45.1801	0.2033	0.0082	0.0030	0.0017	55.0044	0.0	0.5544	0.0079	0.0020	0.0009	9.1872
0.02	11.0	3300	0.3183	66.8740	0.19	0.0080	0.0029	0.0016	50.3100	0.0	0.4822	0.0073	0.0019	0.0009	14.1626
0.0172	12.0	3600	0.3428	66.8740	0.2333	0.0088	0.0031	0.0017	43.5784	0.0	0.5133	0.0076	0.0020	0.0009	10.3941
0.0172	13.0	3900	0.3094	45.1801	0.25	0.0091	0.0032	0.0018	45.2613	0.0	0.4867	0.0074	0.0019	0.0009	11.1576
0.0138	14.0	4200	0.3183	100.0000	0.2767	0.0096	0.0033	0.0018	36.5810	0.0	0.5189	0.0076	0.0020	0.0009	10.1970
0.0114	15.0	4500	0.3215	100.0000	0.28	0.0097	0.0033	0.0018	49.6014	0.0	0.5656	0.0079	0.0020	0.0009	7.8818
0.0114	16.0	4800	0.3028	66.8740	0.27	0.0095	0.0033	0.0018	37.8211	0.0	0.5156	0.0076	0.0020	0.0009	9.8522
0.0086	17.0	5100	0.3086	100.0000	0.27	0.0095	0.0033	0.0018	51.0186	0.0	0.5278	0.0077	0.0020	0.0009	8.6946
0.0086	18.0	5400	0.3301	66.8740	0.22	0.0086	0.0031	0.0017	50.3100	0.0	0.5256	0.0076	0.0020	0.0009	9.2118
0.0078	19.0	5700	0.3219	100.0000	0.2533	0.0092	0.0032	0.0018	41.8069	0.0	0.5256	0.0076	0.0020	0.0009	10.2463
0.0067	20.0	6000	0.3188	100.0000	0.2367	0.0089	0.0032	0.0017	48.0957	0.0	0.4744	0.0073	0.0019	0.0009	12.3153
0.0067	21.0	6300	0.3144	100.0000	0.26	0.0093	0.0033	0.0018	44.5527	0.0	0.5444	0.0078	0.0020	0.0009	8.4729
0.0056	22.0	6600	0.3121	100.0000	0.2667	0.0094	0.0033	0.0018	50.8415	0.0	0.5767	0.0080	0.0021	0.0009	7.2167
0.0056	23.0	6900	0.3140	23.3569	0.24	0.0090	0.0032	0.0017	46.6785	0.0	0.5844	0.0081	0.0021	0.0009	6.8473
0.0043	24.0	7200	0.3258	100.0000	0.2567	0.0093	0.0032	0.0018	44.4641	0.0	0.58	0.0080	0.0021	0.0009	7.6355
0.0039	25.0	7500	0.3072	23.3569	0.2333	0.0088	0.0031	0.0017	48.0071	0.0	0.5667	0.0079	0.0020	0.0009	8.3005
0.0039	26.0	7800	0.3147	23.3569	0.22	0.0086	0.0031	0.0017	48.3614	0.0	0.5689	0.0080	0.0020	0.0009	7.5369
0.0028	27.0	8100	0.3168	35.9304	0.2233	0.0086	0.0031	0.0017	45.9699	0.0	0.5567	0.0079	0.0020	0.0009	8.2020
0.0028	28.0	8400	0.3123	23.3569	0.23	0.0088	0.0031	0.0017	46.1470	0.0	0.5489	0.0078	0.0020	0.0009	8.6207
0.0024	29.0	8700	0.3166	23.3569	0.2333	0.0088	0.0031	0.0017	46.2356	0.0	0.5367	0.0077	0.0020	0.0009	9.6552
0.0025	30.0	9000	0.3170	23.3569	0.2333	0.0088	0.0031	0.0017	46.9442	0.0	0.5367	0.0077	0.0020	0.0009	9.6305

Framework versions

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