exp5_10partition_modelo_asl6000

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: 0.1256
• Model Preparation Time: 0.0033
• Bleu Msl: 0
• Bleu 1 Msl: 0
• Bleu 2 Msl: 0
• Bleu 3 Msl: 0
• Bleu 4 Msl: 0
• Ter Msl: 100
• Bleu Asl: 0
• Bleu 1 Asl: 0.9795
• Bleu 2 Asl: 0.9631
• Bleu 3 Asl: 0.9442
• Bleu 4 Asl: 0.9211
• Ter Asl: 2.4470

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	Model Preparation Time	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	150	0.1646	0.0033	0	0	0	0	0	100	0	0.9507	0.9192	0.8868	0.8511	6.0606
No log	2.0	300	0.1208	0.0033	0	0	0	0	0	100	0	0.9639	0.9411	0.9156	0.8859	4.3607
No log	3.0	450	0.1059	0.0033	0	0	0	0	0	100	0	0.9680	0.9455	0.9203	0.8907	4.0650
0.2631	4.0	600	0.1012	0.0033	0	0	0	0	0	100	0	0.9722	0.9531	0.9320	0.9058	3.4368
0.2631	5.0	750	0.0931	0.0033	0	0	0	0	0	100	0	0.9692	0.9498	0.9276	0.9005	3.9172
0.2631	6.0	900	0.0943	0.0033	0	0	0	0	0	100	0	0.9735	0.9550	0.9340	0.9084	3.1412
0.0342	7.0	1050	0.0988	0.0033	0	0	0	0	0	100	0	0.9732	0.9545	0.9336	0.9085	3.4368
0.0342	8.0	1200	0.0864	0.0033	0	0	0	0	0	100	0	0.9755	0.9577	0.9378	0.9139	3.0673
0.0342	9.0	1350	0.0955	0.0033	0	0	0	0	0	100	0	0.9707	0.9511	0.9288	0.9012	3.4738
0.0163	10.0	1500	0.0852	0.0033	0	0	0	0	0	100	0	0.9728	0.9538	0.9331	0.9089	3.2890
0.0163	11.0	1650	0.0956	0.0033	0	0	0	0	0	100	0	0.9726	0.9539	0.9329	0.9077	3.3629
0.0163	12.0	1800	0.0960	0.0033	0	0	0	0	0	100	0	0.9764	0.9595	0.9402	0.9169	2.8086
0.0163	13.0	1950	0.0969	0.0033	0	0	0	0	0	100	0	0.9745	0.9562	0.9358	0.9111	3.1412
0.0097	14.0	2100	0.0919	0.0033	0	0	0	0	0	100	0	0.9771	0.9598	0.9404	0.9170	2.8825
0.0097	15.0	2250	0.1019	0.0033	0	0	0	0	0	100	0	0.9748	0.9579	0.9386	0.9156	3.0303
0.0097	16.0	2400	0.0946	0.0033	0	0	0	0	0	100	0	0.9704	0.9520	0.9310	0.9057	3.8433
0.0059	17.0	2550	0.0917	0.0033	0	0	0	0	0	100	0	0.9758	0.9590	0.9395	0.9163	3.0303
0.0059	18.0	2700	0.0971	0.0033	0	0	0	0	0	100	0	0.9748	0.9583	0.9396	0.9170	3.2151
0.0059	19.0	2850	0.0893	0.0033	0	0	0	0	0	100	0	0.9764	0.9594	0.9392	0.9152	2.8825
0.0048	20.0	3000	0.0889	0.0033	0	0	0	0	0	100	0	0.9774	0.9606	0.9409	0.9166	2.9194
0.0048	21.0	3150	0.0939	0.0033	0	0	0	0	0	100	0	0.9774	0.9608	0.9414	0.9175	2.7716
0.0048	22.0	3300	0.0931	0.0033	0	0	0	0	0	100	0	0.9771	0.9607	0.9415	0.9177	2.7716
0.0048	23.0	3450	0.0968	0.0033	0	0	0	0	0	100	0	0.9784	0.9625	0.9440	0.9214	2.6238
0.0022	24.0	3600	0.0992	0.0033	0	0	0	0	0	100	0	0.9780	0.9619	0.9429	0.9198	2.6608
0.0022	25.0	3750	0.0957	0.0033	0	0	0	0	0	100	0	0.9777	0.9620	0.9435	0.9210	2.7347
0.0022	26.0	3900	0.0967	0.0033	0	0	0	0	0	100	0	0.9768	0.9605	0.9414	0.9179	2.8086
0.0019	27.0	4050	0.0976	0.0033	0	0	0	0	0	100	0	0.9771	0.9613	0.9426	0.9196	2.7716
0.0019	28.0	4200	0.0950	0.0033	0	0	0	0	0	100	0	0.9781	0.9624	0.9438	0.9207	2.6608
0.0019	29.0	4350	0.0954	0.0033	0	0	0	0	0	100	0	0.9777	0.9618	0.9429	0.9195	2.6977
0.002	30.0	4500	0.0951	0.0033	0	0	0	0	0	100	0	0.9781	0.9622	0.9433	0.9199	2.6238

Framework versions

• Transformers 4.50.2
• Pytorch 2.6.0+cu124
• Datasets 3.5.0
• Tokenizers 0.21.1