9000samples

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: 0.1413
• Model Preparation Time: 0.0237
• Bleu Msl: 91.1702
• Bleu Asl: 0
• Ter Msl: 5.4860
• 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: 1e-05
• train_batch_size: 32
• eval_batch_size: 64
• seed: 42
• optimizer: Use 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 Asl	Ter Msl	Ter Asl
No log	1.0	225	1.6134	0.0237	4.2664	37.3206	123.2008	66.8905
No log	2.0	450	1.0476	0.0237	39.9699	78.0861	46.875	10.4070
1.7524	3.0	675	0.6934	0.0237	56.6935	81.8456	31.1553	8.7285
1.7524	4.0	900	0.5114	0.0237	61.6302	83.6277	26.0417	7.9312
0.6148	5.0	1125	0.4408	0.0237	68.7337	85.4571	20.4545	7.3437
0.6148	6.0	1350	0.3971	0.0237	68.8182	86.5750	20.0758	6.8821
0.3289	7.0	1575	0.3596	0.0237	68.7621	88.0049	20.1705	6.1687
0.3289	8.0	1800	0.3205	0.0237	72.2466	88.9222	17.2348	5.7910
0.2228	9.0	2025	0.2840	0.0237	73.0662	89.4050	16.4773	5.5812
0.2228	10.0	2250	0.2596	0.0237	74.5818	90.3036	15.9091	5.2035
0.2228	11.0	2475	0.2465	0.0237	74.8058	89.9123	15.625	5.3714
0.1439	12.0	2700	0.2316	0.0237	77.3348	89.9976	14.2992	5.2035
0.1439	13.0	2925	0.2236	0.0237	75.9479	90.0635	14.2045	5.2035
0.1078	14.0	3150	0.2204	0.0237	75.6599	90.4806	14.3939	5.1196
0.1078	15.0	3375	0.2188	0.0237	75.8864	90.2107	14.7727	5.2455
0.0862	16.0	3600	0.2111	0.0237	76.4512	90.2883	14.6780	5.2455
0.0862	17.0	3825	0.2049	0.0237	76.5906	90.5898	14.3939	5.0776
0.0744	18.0	4050	0.2047	0.0237	76.1354	90.1256	14.6780	5.2875
0.0744	19.0	4275	0.2011	0.0237	76.4851	90.9593	14.6780	4.9098
0.0648	20.0	4500	0.2002	0.0237	77.1588	91.0477	14.0152	4.8258
0.0648	21.0	4725	0.2003	0.0237	77.6252	90.5716	13.8258	5.1196
0.0648	22.0	4950	0.1980	0.0237	78.6152	90.9392	13.4470	4.9517
0.0575	23.0	5175	0.1966	0.0237	78.5031	90.7533	13.5417	4.9937
0.0575	24.0	5400	0.1962	0.0237	78.5453	90.8408	13.4470	5.0357
0.0531	25.0	5625	0.1952	0.0237	78.3506	90.7533	13.3523	4.9937
0.0531	26.0	5850	0.1938	0.0237	79.0856	90.9085	12.9735	4.9517
0.0492	27.0	6075	0.1930	0.0237	79.0743	90.7829	12.9735	4.9517
0.0492	28.0	6300	0.1928	0.0237	79.8281	90.8176	12.7841	4.9937
0.0465	29.0	6525	0.1929	0.0237	79.4814	90.7628	12.8788	4.9937
0.0465	30.0	6750	0.1929	0.0237	79.3085	90.7628	12.9735	4.9937

Framework versions

• Transformers 4.46.2
• Pytorch 2.5.1+cu121
• Datasets 3.1.0
• Tokenizers 0.20.3