exp5_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: 0.8684
• Model Preparation Time: 0.0035
• Bleu Msl: 0
• Bleu 1 Msl: 0.7062
• Bleu 2 Msl: 0.5800
• Bleu 3 Msl: 0.4740
• Bleu 4 Msl: 0.3480
• Ter Msl: 37.7193
• Bleu Asl: 0
• Bleu 1 Asl: 0.9629
• Bleu 2 Asl: 0.9424
• Bleu 3 Asl: 0.9231
• Bleu 4 Asl: 0.8965
• Ter Asl: 4.0580

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	0.9786	0.0035	0	0.0256	0.0165	0.0107	0.0049	831.3552	0	0.9330	0.8927	0.8514	0.8121	8.2202
No log	2.0	300	0.6154	0.0035	0	0.2507	0.1823	0.1371	0.0933	105.4916	0	0.9532	0.9232	0.8892	0.8542	5.8069
No log	3.0	450	0.5349	0.0035	0	0.7027	0.5836	0.4828	0.3657	29.5837	0	0.9532	0.9231	0.8901	0.8561	5.8069
0.5113	4.0	600	0.5290	0.0035	0	0.6919	0.5959	0.5164	0.4092	33.4810	0	0.9591	0.9322	0.9021	0.8716	5.1282
0.5113	5.0	750	0.5799	0.0035	0	0.7036	0.5717	0.4799	0.3644	37.9982	0	0.9622	0.9363	0.9075	0.8784	5.0528
0.5113	6.0	900	0.6210	0.0035	0	0.675	0.5696	0.4844	0.3783	37.4668	0	0.9615	0.9365	0.9094	0.8802	5.0528
0.0686	7.0	1050	0.6507	0.0035	0	0.6486	0.5360	0.4426	0.3261	40.8326	0	0.9641	0.9389	0.9117	0.8830	4.5249
0.0686	8.0	1200	0.7079	0.0035	0	0.6058	0.4845	0.3913	0.2856	45.9699	0	0.9629	0.9372	0.9089	0.8782	4.6757
0.0686	9.0	1350	0.6309	0.0035	0	0.6959	0.5640	0.4764	0.3747	38.0868	0	0.9622	0.9368	0.9093	0.8802	5.0528
0.0315	10.0	1500	0.6555	0.0035	0	0.6784	0.5393	0.4335	0.3132	40.2126	0	0.9642	0.9386	0.9105	0.8807	4.6757
0.0315	11.0	1650	0.5855	0.0035	0	0.6667	0.5319	0.4366	0.3428	39.0611	0	0.9681	0.9451	0.9190	0.8904	4.2232
0.0315	12.0	1800	0.5110	0.0035	0	0.7456	0.6591	0.5766	0.4690	28.8751	0	0.9642	0.9383	0.9096	0.8794	4.6003
0.0315	13.0	1950	0.6370	0.0035	0	0.7238	0.5590	0.4327	0.3156	37.6439	0	0.9649	0.9407	0.9144	0.8870	4.4495
0.0192	14.0	2100	0.6059	0.0035	0	0.6777	0.5252	0.4378	0.3498	41.6298	0	0.9648	0.9401	0.9128	0.8839	4.6757
0.0192	15.0	2250	0.4800	0.0035	0	0.7336	0.6387	0.5654	0.4748	30.2923	0	0.8313	0.7981	0.7608	0.7157	23.2278
0.0192	16.0	2400	0.5424	0.0035	0	0.7381	0.6389	0.5627	0.4672	29.4951	0	0.9662	0.9425	0.9155	0.8879	4.2232
0.0113	17.0	2550	0.5310	0.0035	0	0.7466	0.6593	0.5865	0.4928	27.7236	0	0.8290	0.7942	0.7569	0.7120	23.5294
0.0113	18.0	2700	0.5556	0.0035	0	0.7440	0.6364	0.5504	0.4503	29.9380	0	0.9616	0.9342	0.9042	0.8726	4.8265
0.0113	19.0	2850	0.5521	0.0035	0	0.7456	0.6411	0.5657	0.4743	29.5837	0	0.9655	0.9409	0.9141	0.8852	4.4495
0.0101	20.0	3000	0.5658	0.0035	0	0.7316	0.6419	0.5665	0.4701	29.4066	0	0.9668	0.9428	0.9150	0.8857	4.2232
0.0101	21.0	3150	0.5093	0.0035	0	0.7331	0.6524	0.5839	0.4941	25.8636	0	0.9648	0.9398	0.9120	0.8826	4.5249
0.0101	22.0	3300	0.5516	0.0035	0	0.7487	0.6693	0.5954	0.5015	27.1922	0	0.9668	0.9432	0.9155	0.8866	4.2986
0.0101	23.0	3450	0.5263	0.0035	0	0.7502	0.6698	0.6003	0.5118	26.1293	0	0.9681	0.9446	0.9174	0.8887	4.2232
0.0069	24.0	3600	0.5287	0.0035	0	0.7502	0.6698	0.5979	0.5051	25.5979	0	0.9648	0.9398	0.9120	0.8830	4.4495
0.0069	25.0	3750	0.5944	0.0035	0	0.7016	0.5979	0.5148	0.4133	33.3924	0	0.9681	0.9450	0.9188	0.8916	4.1478
0.0069	26.0	3900	0.5547	0.0035	0	0.7466	0.6598	0.5809	0.4724	27.9008	0	0.9674	0.9447	0.9190	0.8921	4.2232
0.0051	27.0	4050	0.5918	0.0035	0	0.7451	0.6390	0.5474	0.4306	30.2037	0	0.9681	0.9450	0.9192	0.8922	4.1478
0.0051	28.0	4200	0.5634	0.0035	0	0.7511	0.6606	0.5806	0.4761	28.0779	0	0.9694	0.9469	0.9214	0.8946	3.9970
0.0051	29.0	4350	0.5602	0.0035	0	0.7526	0.6626	0.5838	0.4809	27.9894	0	0.9694	0.9469	0.9214	0.8946	3.9970
0.004	30.0	4500	0.5592	0.0035	0	0.7536	0.6642	0.5862	0.4797	27.9008	0	0.9694	0.9469	0.9214	0.8946	3.9970

Framework versions

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