IAmSkyDra/BARTBana_Translation_InsertBase_v0

BARTBana_Translation_DeleteBase_v0

This model is a fine-tuned version of IAmSkyDra/BARTBana_v4 on the None dataset. It achieves the following results on the evaluation set:

• Loss: 0.4460
• Sacrebleu: 4.8908

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: 2e-05
• train_batch_size: 100
• eval_batch_size: 100
• 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: 15
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Sacrebleu
0.7976	1.0	748	0.7130	2.2026
0.6905	2.0	1496	0.6247	2.8736
0.6189	3.0	2244	0.5805	3.2935
0.5677	4.0	2992	0.5484	3.6212
0.5304	5.0	3740	0.5260	3.9386
0.5018	6.0	4488	0.5084	4.1127
0.4756	7.0	5236	0.4935	4.2759
0.4576	8.0	5984	0.4810	4.4265
0.4336	9.0	6732	0.4720	4.6182
0.422	10.0	7480	0.4645	4.6528
0.4124	11.0	8228	0.4567	4.7800
0.3921	12.0	8976	0.4516	4.7846
0.3894	13.0	9724	0.4479	4.8671
0.3809	14.0	10472	0.4464	4.8673
0.3745	15.0	11220	0.4460	4.8908

Framework versions

• Transformers 4.48.1
• Pytorch 2.5.1+cu124
• Datasets 3.2.0
• Tokenizers 0.21.0