vit5-large_nli

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

• Loss: 1.8846
• Accuracy: 0.8018
• Precision Macro: 0.8019
• Recall Macro: 0.8020
• F1 Macro: 0.8018
• F1 Weighted: 0.8017

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: 64
• eval_batch_size: 64
• seed: 42
• gradient_accumulation_steps: 2
• total_train_batch_size: 128
• optimizer: Use adamw_torch with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: cosine
• lr_scheduler_warmup_ratio: 0.1
• num_epochs: 20
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	Precision Macro	Recall Macro	F1 Macro	F1 Weighted
1.0755	1.0	143	0.7110	0.7020	0.7065	0.7026	0.7008	0.7007
0.5707	2.0	286	0.5831	0.7800	0.7892	0.7793	0.7800	0.7801
0.3356	3.0	429	0.6000	0.7911	0.7927	0.7916	0.7910	0.7909
0.1497	4.0	572	0.7687	0.7827	0.7848	0.7830	0.7826	0.7825
0.0876	5.0	715	0.8672	0.7867	0.7892	0.7864	0.7868	0.7868
0.0601	6.0	858	1.1073	0.7863	0.7869	0.7862	0.7862	0.7862
0.0473	7.0	1001	1.2264	0.7769	0.7821	0.7777	0.7762	0.7760
0.037	8.0	1144	1.1917	0.7947	0.7956	0.7945	0.7948	0.7948
0.0256	9.0	1287	1.3581	0.7867	0.7869	0.7866	0.7867	0.7867
0.0188	10.0	1430	1.3638	0.7916	0.7919	0.7916	0.7915	0.7916
0.0153	11.0	1573	1.5960	0.7902	0.7914	0.7903	0.7903	0.7903
0.0101	12.0	1716	1.6123	0.7938	0.7938	0.7939	0.7938	0.7937
0.0098	13.0	1859	1.7553	0.8	0.8017	0.8004	0.7999	0.7999
0.006	14.0	2002	1.7906	0.7978	0.7985	0.7982	0.7977	0.7975
0.0047	15.0	2145	1.8154	0.7991	0.7992	0.7993	0.7991	0.7991
0.0034	16.0	2288	1.8285	0.8013	0.8015	0.8016	0.8013	0.8012
0.0018	17.0	2431	1.8543	0.8004	0.8006	0.8007	0.8004	0.8003
0.0021	18.0	2574	1.8807	0.8018	0.8019	0.8020	0.8018	0.8017
0.0009	19.0	2717	1.8842	0.8013	0.8014	0.8015	0.8013	0.8013
0.0019	20.0	2860	1.8846	0.8018	0.8019	0.8020	0.8018	0.8017

Framework versions

• Transformers 4.55.0
• Pytorch 2.7.0+cu126
• Datasets 4.0.0
• Tokenizers 0.21.4