exp2_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: 0.3430
• Model Preparation Time: 0.0033
• Bleu Msl: 0
• Bleu 1 Msl: 0.8277
• Bleu 2 Msl: 0.7696
• Bleu 3 Msl: 0.7052
• Bleu 4 Msl: 0.5772
• Ter Msl: 22.5957
• Bleu Asl: 0
• Bleu 1 Asl: 0.9456
• Bleu 2 Asl: 0.9182
• Bleu 3 Asl: 0.8897
• Bleu 4 Asl: 0.8582
• Ter Asl: 7.0887

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	0.3065	0.0033	0	0.7311	0.6296	0.5211	0.3540	27.2443	0	0.9288	0.8956	0.8610	0.8204	9.2642
0.4646	2.0	600	0.2634	0.0033	0	0.4298	0.3458	0.2674	0.1713	33.1942	0	0.8522	0.8164	0.7786	0.7315	19.5657
0.4646	3.0	900	0.2522	0.0033	0	0.2858	0.2266	0.1703	0.1032	30.0626	0	0.8462	0.8119	0.7755	0.7320	20.1206
0.1011	4.0	1200	0.2632	0.0033	0	0.8079	0.7069	0.5981	0.4302	26.2004	0	0.8595	0.8280	0.7942	0.7509	6.4174
0.0585	5.0	1500	0.2415	0.0033	0	0.7160	0.6243	0.5197	0.3395	35.8038	0	0.9394	0.9155	0.8900	0.8588	7.4065
0.0585	6.0	1800	0.2693	0.0033	0	0.7674	0.6792	0.5818	0.4110	29.1232	0	0.9439	0.9128	0.8807	0.8454	7.4065
0.0371	7.0	2100	0.2550	0.0033	0	0.8081	0.7227	0.6181	0.4363	27.0355	0	0.9332	0.9066	0.8783	0.8441	9.9397
0.0371	8.0	2400	0.2676	0.0033	0	0.7817	0.7014	0.6085	0.4450	29.1232	0	0.9169	0.8799	0.8431	0.8006	10.7841
0.0252	9.0	2700	0.2742	0.0033	0	0.8216	0.7296	0.6202	0.4558	28.9144	0	0.9499	0.9267	0.9020	0.8715	6.4897
0.0203	10.0	3000	0.2563	0.0033	0	0.7882	0.7013	0.5895	0.4080	29.9582	0	0.9265	0.9016	0.8749	0.8428	9.0712
0.0203	11.0	3300	0.2492	0.0033	0	0.8598	0.7997	0.7090	0.5392	20.2505	0	0.9336	0.9060	0.8778	0.8447	7.4789
0.0165	12.0	3600	0.2599	0.0033	0	0.8214	0.7465	0.6463	0.4619	27.1399	0	0.9192	0.8949	0.8689	0.8362	10.4704
0.0165	13.0	3900	0.2675	0.0033	0	0.7345	0.6490	0.5383	0.3694	34.5511	0	0.9280	0.9020	0.8755	0.8435	9.2159
0.0144	14.0	4200	0.2876	0.0033	0	0.7839	0.7050	0.6018	0.4132	31.1065	0	0.9339	0.9086	0.8829	0.8511	8.4439
0.012	15.0	4500	0.2766	0.0033	0	0.7646	0.6795	0.5634	0.3696	31.4196	0	0.9371	0.9155	0.8916	0.8614	8.9023
0.012	16.0	4800	0.3112	0.0033	0	0.7465	0.6584	0.5477	0.3614	35.3862	0	0.9433	0.9168	0.8902	0.8578	7.2135
0.0095	17.0	5100	0.2828	0.0033	0	0.7348	0.6479	0.5411	0.3658	34.4468	0	0.9345	0.9123	0.8874	0.8562	9.2642
0.0095	18.0	5400	0.2860	0.0033	0	0.7813	0.7058	0.6101	0.4214	29.2276	0	0.9350	0.9035	0.8726	0.8367	9.0953
0.008	19.0	5700	0.3017	0.0033	0	0.7588	0.6723	0.5606	0.3619	32.0459	0	0.9331	0.9045	0.8766	0.8428	9.1435
0.0063	20.0	6000	0.3017	0.0033	0	0.7835	0.6994	0.5861	0.3939	31.1065	0	0.9467	0.9185	0.8903	0.8564	6.8034
0.0063	21.0	6300	0.2927	0.0033	0	0.7815	0.6990	0.5881	0.3950	29.6451	0	0.9461	0.9182	0.8900	0.8568	7.2135
0.0056	22.0	6600	0.2882	0.0033	0	0.8091	0.7331	0.6307	0.4426	26.9311	0	0.9397	0.9115	0.8822	0.8483	8.2027
0.0056	23.0	6900	0.2905	0.0033	0	0.7966	0.7133	0.6112	0.4232	29.3319	0	0.9386	0.9095	0.8805	0.8467	8.0820
0.0053	24.0	7200	0.2848	0.0033	0	0.8024	0.7242	0.6236	0.4325	28.0793	0	0.9327	0.9022	0.8727	0.8384	9.0470
0.0043	25.0	7500	0.2843	0.0033	0	0.8069	0.7282	0.6233	0.4347	27.4530	0	0.9341	0.9026	0.8719	0.8372	8.3474
0.0043	26.0	7800	0.2807	0.0033	0	0.8200	0.7516	0.6519	0.4648	25.2610	0	0.9442	0.9147	0.8862	0.8529	7.1894
0.0034	27.0	8100	0.2833	0.0033	0	0.8163	0.7447	0.6471	0.4606	25.2610	0	0.9346	0.9035	0.8726	0.8364	8.6369
0.0034	28.0	8400	0.2802	0.0033	0	0.8034	0.7274	0.6196	0.4281	27.2443	0	0.9336	0.9033	0.8734	0.8380	8.8782
0.0033	29.0	8700	0.2786	0.0033	0	0.7994	0.7223	0.6142	0.4228	27.2443	0	0.9331	0.9017	0.8710	0.8348	8.8299
0.0025	30.0	9000	0.2792	0.0033	0	0.7974	0.7215	0.6136	0.4230	27.3486	0	0.9340	0.9031	0.8727	0.8372	8.6852

Framework versions

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