exp5_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: 1.5231
• Bleu Msl: 100.0000
• Bleu 1 Msl: 0.39
• Bleu 2 Msl: 0.0114
• Bleu 3 Msl: 0.0037
• Bleu 4 Msl: 0.0020
• Ter Msl: 37.0614
• 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	1.6460	66.8740	0.0467	0.0040	0.0019	0.0012	1073.6935	0	0	0	0	0	100
No log	2.0	150	1.0511	100.0000	0.2433	0.0090	0.0032	0.0017	134.3667	0	0	0	0	0	100
No log	3.0	225	1.0074	100.0000	0.3	0.0100	0.0034	0.0018	36.3153	0	0	0	0	0	100
No log	4.0	300	1.2529	100.0000	0.2467	0.0091	0.0032	0.0017	42.4269	0	0	0	0	0	100
No log	5.0	375	0.8305	100.0000	0.4133	0.0118	0.0038	0.0020	28.0779	0	0	0	0	0	100
No log	6.0	450	1.1453	100.0000	0.2633	0.0094	0.0033	0.0018	39.4154	0	0	0	0	0	100
0.566	7.0	525	1.0225	100.0000	0.3567	0.0109	0.0036	0.0019	34.0124	0	0	0	0	0	100
0.566	8.0	600	1.1880	100.0000	0.3033	0.0101	0.0034	0.0018	37.1125	0	0	0	0	0	100
0.566	9.0	675	1.2738	100.0000	0.3167	0.0103	0.0035	0.0019	33.3038	0	0	0	0	0	100
0.566	10.0	750	1.3089	100.0000	0.2433	0.0090	0.0032	0.0017	45.6156	0	0	0	0	0	100
0.566	11.0	825	0.9362	100.0000	0.3	0.0100	0.0034	0.0018	33.2152	0	0	0	0	0	100
0.566	12.0	900	1.4018	100.0000	0.2633	0.0094	0.0033	0.0018	41.7183	0	0	0	0	0	100
0.566	13.0	975	1.3847	100.0000	0.3067	0.0101	0.0034	0.0018	39.4154	0	0	0	0	0	100
0.0427	14.0	1050	1.1430	100.0000	0.2867	0.0098	0.0034	0.0018	34.5438	0	0	0	0	0	100
0.0427	15.0	1125	1.1585	100.0000	0.3133	0.0102	0.0035	0.0019	33.4810	0	0	0	0	0	100
0.0427	16.0	1200	1.2583	100.0000	0.32	0.0103	0.0035	0.0019	35.1639	0	0	0	0	0	100
0.0427	17.0	1275	1.3332	100.0000	0.26	0.0093	0.0033	0.0018	40.2126	0	0	0	0	0	100
0.0427	18.0	1350	1.4169	100.0000	0.3	0.0100	0.0034	0.0018	38.4411	0	0	0	0	0	100
0.0427	19.0	1425	1.2117	100.0000	0.3067	0.0101	0.0034	0.0018	35.2524	0	0	0	0	0	100
0.0166	20.0	1500	1.2812	100.0000	0.3033	0.0101	0.0034	0.0018	36.2267	0	0	0	0	0	100
0.0166	21.0	1575	1.3728	100.0000	0.2833	0.0097	0.0034	0.0018	36.7582	0	0	0	0	0	100
0.0166	22.0	1650	1.3649	100.0000	0.28	0.0097	0.0033	0.0018	35.9610	0	0	0	0	0	100
0.0166	23.0	1725	1.3372	100.0000	0.2933	0.0099	0.0034	0.0018	35.1639	0	0	0	0	0	100
0.0166	24.0	1800	1.3640	100.0000	0.2933	0.0099	0.0034	0.0018	37.7325	0	0	0	0	0	100
0.0166	25.0	1875	1.3624	100.0000	0.2733	0.0096	0.0033	0.0018	36.5810	0	0	0	0	0	100
0.0166	26.0	1950	1.3747	100.0000	0.28	0.0097	0.0033	0.0018	36.5810	0	0	0	0	0	100
0.0103	27.0	2025	1.3498	100.0000	0.2967	0.0100	0.0034	0.0018	36.9353	0	0	0	0	0	100
0.0103	28.0	2100	1.3475	100.0000	0.2967	0.0100	0.0034	0.0018	36.1382	0	0	0	0	0	100
0.0103	29.0	2175	1.3758	100.0000	0.2967	0.0100	0.0034	0.0018	36.1382	0	0	0	0	0	100
0.0103	30.0	2250	1.3763	100.0000	0.2933	0.0099	0.0034	0.0018	36.3153	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