exp3_10partition_modelo6000

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: 0.5689
• Model Preparation Time: 0.0034
• Bleu Msl: 0
• Bleu 1 Msl: 0.8320
• Bleu 2 Msl: 0.7371
• Bleu 3 Msl: 0.6111
• Bleu 4 Msl: 0.4747
• Ter Msl: 25.5682
• Bleu Asl: 0
• Bleu 1 Asl: 0.9731
• Bleu 2 Asl: 0.9522
• Bleu 3 Asl: 0.9291
• Bleu 4 Asl: 0.9010
• Ter Asl: 3.0735

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	1.3308	0.0034	0	0.5309	0.3589	0.2253	0.1104	63.2378	0	0.9398	0.9034	0.8665	0.8255	7.3964
No log	2.0	300	1.1970	0.0034	0	0.2508	0.1620	0.1042	0.0581	112.1417	0	0.9605	0.9342	0.9066	0.8729	4.8817
No log	3.0	450	1.1426	0.0034	0	0.1355	0.0829	0.0489	0.0249	195.9528	0	0.9644	0.9381	0.9099	0.8770	4.3639
0.5052	4.0	600	1.1307	0.0034	0	0.4226	0.2837	0.1899	0.1238	78.3305	0	0.9651	0.9391	0.9109	0.8784	4.3639
0.5052	5.0	750	1.0743	0.0034	0	0.6482	0.4648	0.3387	0.2374	56.4081	0	0.9677	0.9436	0.9185	0.8884	3.9201
0.5052	6.0	900	1.0719	0.0034	0	0.6376	0.4993	0.3970	0.2993	47.3862	0	0.9651	0.9395	0.9130	0.8822	4.4379
0.0684	7.0	1050	1.1431	0.0034	0	0.6444	0.4855	0.3685	0.2702	50.5902	0	0.9644	0.9393	0.9130	0.8817	4.1420
0.0684	8.0	1200	1.1090	0.0034	0	0.6562	0.4803	0.3638	0.2656	51.1804	0	0.7332	0.6959	0.6578	0.6083	3.9201
0.0684	9.0	1350	1.1236	0.0034	0	0.6855	0.5231	0.4066	0.3001	49.0725	0	0.9664	0.9425	0.9171	0.8874	3.7722
0.031	10.0	1500	1.1382	0.0034	0	0.6427	0.4649	0.3347	0.2369	55.9865	0	0.9638	0.9383	0.9113	0.8786	4.2160
0.031	11.0	1650	1.1146	0.0034	0	0.6577	0.4821	0.3636	0.2490	51.0118	0	0.9651	0.9412	0.9151	0.8840	3.9201
0.031	12.0	1800	1.1889	0.0034	0	0.6732	0.5063	0.3848	0.2800	52.0236	0	0.9651	0.9410	0.9150	0.8839	3.9201
0.031	13.0	1950	1.1853	0.0034	0	0.6673	0.4892	0.3712	0.2689	52.8668	0	0.9696	0.9479	0.9242	0.8957	3.6243
0.0208	14.0	2100	1.2047	0.0034	0	0.6737	0.5096	0.3857	0.2800	50.3373	0	0.9702	0.9477	0.9242	0.8954	3.6243
0.0208	15.0	2250	1.1953	0.0034	0	0.6619	0.4949	0.3772	0.2729	51.4334	0	0.9657	0.9426	0.9168	0.8865	3.9201
0.0208	16.0	2400	1.1772	0.0034	0	0.6871	0.5183	0.4026	0.2952	48.9882	0	0.9696	0.9474	0.9240	0.8957	3.4763
0.014	17.0	2550	1.1982	0.0034	0	0.6947	0.5251	0.4045	0.2893	47.0489	0	0.9709	0.9502	0.9277	0.9002	3.3284
0.014	18.0	2700	1.2218	0.0034	0	0.6736	0.5080	0.3875	0.2808	48.9039	0	0.9715	0.9515	0.9294	0.9022	3.4024
0.014	19.0	2850	1.2035	0.0034	0	0.6864	0.5211	0.4033	0.2935	49.8314	0	0.9702	0.9493	0.9263	0.8986	3.4763
0.0092	20.0	3000	1.2138	0.0034	0	0.6824	0.5164	0.3985	0.2927	48.0607	0	0.9722	0.9520	0.9299	0.9030	3.1805
0.0092	21.0	3150	1.2348	0.0034	0	0.6718	0.5070	0.3923	0.2902	50.1686	0	0.9696	0.9470	0.9226	0.8936	3.5503
0.0092	22.0	3300	1.2350	0.0034	0	0.6721	0.5071	0.3915	0.2894	49.4941	0	0.9683	0.9457	0.9211	0.8919	3.6243
0.0092	23.0	3450	1.2337	0.0034	0	0.6900	0.5233	0.4046	0.2993	48.4823	0	0.9702	0.9489	0.9256	0.8972	3.5503
0.0075	24.0	3600	1.2332	0.0034	0	0.6919	0.5196	0.3982	0.2903	50.6745	0	0.9702	0.9493	0.9263	0.8983	3.4763
0.0075	25.0	3750	1.2324	0.0034	0	0.6989	0.5313	0.4142	0.3059	48.8196	0	0.9670	0.9455	0.9221	0.8942	3.8462
0.0075	26.0	3900	1.2377	0.0034	0	0.6982	0.5309	0.4119	0.3025	48.3980	0	0.9683	0.9478	0.9253	0.8980	3.6982
0.0068	27.0	4050	1.2398	0.0034	0	0.6893	0.5225	0.4056	0.2996	49.9157	0	0.9677	0.9467	0.9235	0.8956	3.7722
0.0068	28.0	4200	1.2441	0.0034	0	0.6850	0.5211	0.4055	0.3016	49.4098	0	0.9677	0.9467	0.9239	0.8962	3.7722
0.0068	29.0	4350	1.2470	0.0034	0	0.6817	0.5178	0.4043	0.3013	49.2411	0	0.9683	0.9478	0.9253	0.8980	3.6982
0.0051	30.0	4500	1.2511	0.0034	0	0.6795	0.5164	0.4029	0.3006	49.4941	0	0.9683	0.9478	0.9253	0.8980	3.6982

Framework versions

• Transformers 4.50.0
• Pytorch 2.6.0+cu124
• Datasets 3.4.1
• Tokenizers 0.21.1