mbart50-es-zapv3

This model is a fine-tuned version of facebook/mbart-large-50 on the None dataset. It achieves the following results on the evaluation set:

• Loss: 0.2895
• Bleu: 23.2680
• Ter: 62.9771
• Meteor: 0.5012
• Chrf: 49.1953

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: 3e-05
• train_batch_size: 8
• eval_batch_size: 8
• seed: 42
• optimizer: Use OptimizerNames.ADAMW_TORCH_FUSED with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: linear
• num_epochs: 50
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Bleu	Ter	Meteor	Chrf
6.1208	1.0	179	0.4628	7.8281	84.4466	0.2642	30.7944
0.7001	2.0	358	0.6760	0.0	100.0	0.0	0.0074
0.2437	3.0	537	0.2388	16.4870	72.8053	0.3896	40.2706
0.1298	4.0	716	0.2254	19.4490	70.2290	0.4335	44.2058
0.0853	5.0	895	0.2333	18.8371	66.0305	0.4509	44.7725
0.0722	6.0	1074	0.2472	23.2294	63.8359	0.4851	47.3178
0.0389	7.0	1253	0.2645	21.4190	64.5038	0.4614	46.9929
0.0404	8.0	1432	0.2661	20.9712	64.9809	0.4664	47.0832
0.0299	9.0	1611	0.2761	22.9324	64.4084	0.4989	49.5780
0.0239	10.0	1790	0.2821	22.7716	62.7863	0.4986	49.8536
0.0261	11.0	1969	0.2808	24.7373	60.3053	0.5092	50.6788
0.0165	12.0	2148	0.2866	22.7118	60.3053	0.5112	50.6475
0.0175	13.0	2327	0.3255	19.5063	64.4084	0.4480	46.9423
0.0161	14.0	2506	0.2953	24.3036	60.8779	0.5021	50.7556
0.0150	15.0	2685	0.2917	25.2649	60.6870	0.5010	49.9172
0.0193	16.0	2864	0.2876	22.9039	59.7328	0.5004	50.0787
0.0315	17.0	3043	0.2895	23.2680	62.9771	0.5012	49.1953

Framework versions

• Transformers 5.5.3
• Pytorch 2.10.0+cu128
• Datasets 4.8.4
• Tokenizers 0.22.2