exp4_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: 0.5006
• Model Preparation Time: 0.0033
• Bleu Msl: 0
• Bleu 1 Msl: 0.6440
• Bleu 2 Msl: 0.5227
• Bleu 3 Msl: 0.3812
• Bleu 4 Msl: 0.2348
• Ter Msl: 40.3766
• Bleu Asl: 0
• Bleu 1 Asl: 0.9510
• Bleu 2 Asl: 0.9242
• Bleu 3 Asl: 0.8993
• Bleu 4 Asl: 0.8711
• Ter Asl: 6.3438

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	0.4928	0.0033	0	0.0417	0.0283	0.0188	0.0098	492.6755	0	0.8798	0.8450	0.8094	0.7661	10.5189
0.4412	2.0	600	0.4181	0.0033	0	0.2268	0.1694	0.1183	0.0696	96.7447	0	0.9379	0.9123	0.8852	0.8530	8.2179
0.4412	3.0	900	0.4104	0.0033	0	0.6755	0.5733	0.4482	0.2823	38.8606	0	0.9535	0.9279	0.9000	0.8671	6.3865
0.0907	4.0	1200	0.3944	0.0033	0	0.4412	0.3507	0.2565	0.1617	46.5921	0	0.9465	0.9253	0.9022	0.8739	7.4665
0.0538	5.0	1500	0.3892	0.0033	0	0.6268	0.5198	0.4052	0.2851	42.9298	0	0.9443	0.9215	0.8979	0.8690	7.5605
0.0538	6.0	1800	0.4045	0.0033	0	0.6781	0.5738	0.4447	0.3044	44.0488	0	0.9338	0.9100	0.8846	0.8538	8.9692
0.0338	7.0	2100	0.3856	0.0033	0	0.6892	0.6006	0.4937	0.3532	38.5554	0	0.9525	0.9320	0.9108	0.8850	6.2691
0.0338	8.0	2400	0.4031	0.0033	0	0.6096	0.5161	0.4153	0.3011	45.2696	0	0.9483	0.9262	0.9043	0.8777	6.6448
0.027	9.0	2700	0.4079	0.0033	0	0.6236	0.5318	0.4256	0.3067	45.5748	0	0.8982	0.8708	0.8430	0.8067	13.9235
0.0197	10.0	3000	0.4473	0.0033	0	0.6349	0.5521	0.4440	0.3142	43.2350	0	0.9437	0.9189	0.8936	0.8629	7.2083
0.0197	11.0	3300	0.4227	0.0033	0	0.6922	0.6020	0.4873	0.3465	39.8779	0	0.9436	0.9166	0.8914	0.8621	8.1005
0.0153	12.0	3600	0.4301	0.0033	0	0.6903	0.5910	0.4708	0.3348	40.9969	0	0.9403	0.9145	0.8902	0.8613	7.8657
0.0153	13.0	3900	0.4557	0.0033	0	0.6642	0.5795	0.4612	0.3261	42.2177	0	0.9446	0.9160	0.8876	0.8560	7.1378
0.013	14.0	4200	0.4587	0.0033	0	0.6846	0.6014	0.4900	0.3449	38.6572	0	0.9437	0.9167	0.8910	0.8615	8.1240
0.0115	15.0	4500	0.4284	0.0033	0	0.6066	0.5162	0.4099	0.2923	44.2523	0	0.9170	0.8895	0.8634	0.8326	10.9885
0.0115	16.0	4800	0.4278	0.0033	0	0.6444	0.5534	0.4448	0.3171	41.4039	0	0.9357	0.9096	0.8827	0.8512	8.1944
0.0082	17.0	5100	0.4591	0.0033	0	0.6423	0.5454	0.4356	0.3022	43.4385	0	0.9499	0.9263	0.9027	0.8758	6.6448
0.0082	18.0	5400	0.4547	0.0033	0	0.5884	0.5052	0.4046	0.2833	45.4730	0	0.9461	0.9220	0.8968	0.8667	6.7622
0.0083	19.0	5700	0.4398	0.0033	0	0.6364	0.5390	0.4338	0.3136	43.0315	0	0.9375	0.9120	0.8870	0.8570	7.7718
0.006	20.0	6000	0.4413	0.0033	0	0.6197	0.5291	0.4219	0.3029	44.1506	0	0.9308	0.9092	0.8870	0.8593	8.9692
0.006	21.0	6300	0.4482	0.0033	0	0.6453	0.5521	0.4445	0.3170	42.6246	0	0.9314	0.9101	0.8881	0.8615	8.8518
0.0059	22.0	6600	0.4511	0.0033	0	0.6409	0.5444	0.4389	0.3161	43.0315	0	0.9391	0.9207	0.9011	0.8762	7.9127
0.0059	23.0	6900	0.4328	0.0033	0	0.6444	0.5509	0.4457	0.3234	42.1160	0	0.9464	0.9264	0.9062	0.8814	6.8795
0.0052	24.0	7200	0.4306	0.0033	0	0.6259	0.5366	0.4350	0.3141	44.1506	0	0.9435	0.9249	0.9057	0.8818	7.3022
0.004	25.0	7500	0.4461	0.0033	0	0.6389	0.5488	0.4440	0.3174	42.9298	0	0.9441	0.9252	0.9054	0.8812	7.1613
0.004	26.0	7800	0.4407	0.0033	0	0.6014	0.5101	0.4056	0.2880	45.3713	0	0.9508	0.9281	0.9058	0.8800	6.1282
0.0033	27.0	8100	0.4504	0.0033	0	0.6305	0.5377	0.4321	0.3049	44.3540	0	0.9455	0.9258	0.9056	0.8814	6.7856
0.0033	28.0	8400	0.4467	0.0033	0	0.6404	0.5525	0.4463	0.3165	42.5229	0	0.9494	0.9269	0.9051	0.8798	6.4804
0.0031	29.0	8700	0.4486	0.0033	0	0.6494	0.5589	0.4519	0.3214	42.2177	0	0.9519	0.9304	0.9095	0.8847	6.1752
0.0024	30.0	9000	0.4491	0.0033	0	0.6532	0.5639	0.4561	0.3236	41.8108	0	0.9516	0.9300	0.9090	0.8843	6.1752

Framework versions

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