exp3_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: 0.3817
• Model Preparation Time: 0.0034
• Bleu Msl: 0
• Bleu 1 Msl: 0.7804
• Bleu 2 Msl: 0.6823
• Bleu 3 Msl: 0.5551
• Bleu 4 Msl: 0.4249
• Ter Msl: 33.6174
• Bleu Asl: 0
• Bleu 1 Asl: 0.9531
• Bleu 2 Asl: 0.9311
• Bleu 3 Asl: 0.9083
• Bleu 4 Asl: 0.8815
• Ter Asl: 6.1764

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	300	0.6646	0.0034	0	0.0312	0.0161	0.0082	0.0037	648.9882	0	0.9510	0.9235	0.8931	0.8592	6.3634
0.4437	2.0	600	0.6571	0.0034	0	0.4202	0.2571	0.1500	0.0769	82.7150	0	0.9294	0.9037	0.8753	0.8435	8.3716
0.4437	3.0	900	0.6110	0.0034	0	0.5792	0.3846	0.2524	0.1452	61.4671	0	0.8996	0.8695	0.8393	0.8017	6.8715
0.0963	4.0	1200	0.5833	0.0034	0	0.5626	0.3791	0.2495	0.1544	70.3204	0	0.9314	0.9020	0.8705	0.8340	9.2911
0.055	5.0	1500	0.6079	0.0034	0	0.6084	0.4374	0.3042	0.1969	58.0944	0	0.9647	0.9420	0.9161	0.8872	5.0569
0.055	6.0	1800	0.6174	0.0034	0	0.5749	0.4031	0.2785	0.1668	60.6239	0	0.9603	0.9366	0.9108	0.8814	5.5408
0.0355	7.0	2100	0.5909	0.0034	0	0.3742	0.2703	0.1981	0.1379	86.3406	0	0.9321	0.9099	0.8850	0.8559	8.4926
0.0355	8.0	2400	0.6109	0.0034	0	0.5499	0.3849	0.2736	0.1872	72.1754	0	0.9257	0.9012	0.8726	0.8389	9.9444
0.0273	9.0	2700	0.6152	0.0034	0	0.4619	0.3279	0.2365	0.1651	73.6931	0	0.9344	0.9120	0.8858	0.8531	8.8556
0.0192	10.0	3000	0.6463	0.0034	0	0.6577	0.4867	0.3586	0.2258	52.0236	0	0.9503	0.9293	0.9065	0.8785	7.0409
0.0192	11.0	3300	0.6566	0.0034	0	0.6399	0.4874	0.3749	0.2675	55.1433	0	0.9442	0.9199	0.8933	0.8610	7.3312
0.016	12.0	3600	0.6420	0.0034	0	0.6451	0.5071	0.3992	0.2958	51.2648	0	0.9516	0.9305	0.9072	0.8788	6.5812
0.016	13.0	3900	0.6414	0.0034	0	0.6098	0.4737	0.3643	0.2756	54.3845	0	0.9544	0.9320	0.9079	0.8793	6.4844
0.0142	14.0	4200	0.6417	0.0034	0	0.6099	0.4690	0.3572	0.2654	55.9865	0	0.9484	0.9249	0.9002	0.8710	7.1861
0.0107	15.0	4500	0.6349	0.0034	0	0.6362	0.4884	0.3773	0.2859	53.8786	0	0.9610	0.9400	0.9178	0.8910	5.4440
0.0107	16.0	4800	0.6363	0.0034	0	0.6029	0.4573	0.3366	0.2252	53.5413	0	0.9403	0.9205	0.8981	0.8701	8.0813
0.0088	17.0	5100	0.6230	0.0034	0	0.6437	0.4917	0.3669	0.2546	52.4452	0	0.9543	0.9356	0.9144	0.8886	6.5570
0.0088	18.0	5400	0.6489	0.0034	0	0.6191	0.4778	0.3523	0.2405	54.7218	0	0.9511	0.9328	0.9125	0.8875	6.6054
0.0074	19.0	5700	0.6381	0.0034	0	0.6201	0.4775	0.3588	0.2514	55.1433	0	0.9382	0.9204	0.8988	0.8720	8.5168
0.0065	20.0	6000	0.6152	0.0034	0	0.6278	0.4971	0.3948	0.3076	51.7707	0	0.9474	0.9268	0.9046	0.8778	6.7747
0.0065	21.0	6300	0.6253	0.0034	0	0.6621	0.5167	0.4071	0.3058	51.6020	0	0.9567	0.9384	0.9184	0.8932	5.7101
0.0057	22.0	6600	0.6376	0.0034	0	0.6503	0.5156	0.4066	0.3005	50.0	0	0.9537	0.9361	0.9158	0.8906	6.5328
0.0057	23.0	6900	0.6378	0.0034	0	0.6680	0.5163	0.3966	0.2829	54.8904	0	0.9556	0.9379	0.9172	0.8926	6.4844
0.0045	24.0	7200	0.6320	0.0034	0	0.6645	0.5203	0.4034	0.2947	50.5902	0	0.9508	0.9324	0.9111	0.8850	6.7022
0.0042	25.0	7500	0.6383	0.0034	0	0.6665	0.5116	0.3933	0.2844	51.4334	0	0.9482	0.9290	0.9073	0.8813	6.8473
0.0042	26.0	7800	0.6403	0.0034	0	0.6554	0.5060	0.3886	0.2775	51.6020	0	0.9548	0.9358	0.9144	0.8888	6.2182
0.0035	27.0	8100	0.6438	0.0034	0	0.6640	0.5096	0.3866	0.2699	50.2530	0	0.9594	0.9410	0.9203	0.8954	5.2020
0.0035	28.0	8400	0.6459	0.0034	0	0.6655	0.5088	0.3865	0.2717	51.2648	0	0.9580	0.9398	0.9197	0.8956	5.5408
0.0029	29.0	8700	0.6455	0.0034	0	0.6625	0.5096	0.3918	0.2803	51.3491	0	0.9591	0.9408	0.9203	0.8956	5.4682
0.0028	30.0	9000	0.6424	0.0034	0	0.6659	0.5118	0.3931	0.2808	51.6020	0	0.9587	0.9401	0.9193	0.8944	5.4924

Framework versions

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