vlsp-vit5-15epoch

This model is a fine-tuned version of VietAI/vit5-base on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 0.2038
• Exact Match: 0.852

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: 16
• eval_batch_size: 16
• 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	Exact Match
0.8744	1.0	188	0.6756	0.006
0.605	2.0	376	0.5266	0.088
0.4382	3.0	564	0.4085	0.166
0.3484	4.0	752	0.3264	0.262
0.2782	5.0	940	0.2722	0.414
0.2305	6.0	1128	0.2469	0.52
0.1977	7.0	1316	0.2349	0.568
0.1621	8.0	1504	0.2221	0.642
0.1403	9.0	1692	0.2148	0.634
0.1225	10.0	1880	0.2099	0.74
0.1035	11.0	2068	0.2054	0.75
0.0904	12.0	2256	0.2229	0.728
0.0772	13.0	2444	0.2126	0.776
0.0699	14.0	2632	0.2092	0.84
0.0604	15.0	2820	0.2038	0.852

Framework versions

• Transformers 4.52.4
• Pytorch 2.6.0+cu124
• Datasets 3.6.0
• Tokenizers 0.21.2