vit5-finetuned-vietnamese

This model is a fine-tuned version of VietAI/vit5-base-vietnews-summarization on the None dataset. It achieves the following results on the evaluation set:

• Loss: 0.3329
• Rouge1: 61.9
• Rouge2: 50.51
• Rougel: 55.7
• Rougelsum: 55.62

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: 5e-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: 5
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Rouge1	Rouge2	Rougel	Rougelsum
0.3847	0.5305	200	0.2717	61.27	49.26	54.8	54.79
0.2647	1.0610	400	0.2773	62.02	50.1	55.55	55.49
0.2434	1.5915	600	0.2792	61.86	49.86	55.34	55.34
0.1381	2.1220	800	0.2991	61.76	49.71	55.29	55.26
0.1344	2.6525	1000	0.3012	61.4	49.52	55.07	55.04
0.0917	3.1830	1200	0.3177	61.36	49.89	55.28	55.27
0.0922	3.7135	1400	0.3238	61.81	49.85	55.43	55.38
0.0617	4.2440	1600	0.3280	61.85	50.29	55.51	55.45
0.0567	4.7745	1800	0.3329	61.9	50.51	55.7	55.62

Framework versions

• Transformers 4.57.1
• Pytorch 2.8.0+cu126
• Datasets 4.0.0
• Tokenizers 0.22.1