Revision_PhoBert_Lexical_MetaXLM_relabel

This model is a fine-tuned version of ttqdunggg/Revision_PhoBert_Lexical_Dataset_46k on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 0.5761
• Accuracy: 0.8433
• F1: 0.8392

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: 32
• eval_batch_size: 32
• 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

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	F1
No log	0.2778	100	0.3147	0.8665	0.8609
No log	0.5556	200	0.3097	0.8831	0.8755
No log	0.8333	300	0.3134	0.8760	0.8695
0.2496	1.1111	400	0.3356	0.8611	0.8554
0.2496	1.3889	500	0.3321	0.8777	0.8704
0.2496	1.6667	600	0.3354	0.8678	0.8621
0.2496	1.9444	700	0.3196	0.8832	0.8770
0.1744	2.2222	800	0.4563	0.8485	0.8443
0.1744	2.5	900	0.4871	0.8415	0.8374
0.1744	2.7778	1000	0.5115	0.8360	0.8318
0.1243	3.0556	1100	0.5060	0.8406	0.8364
0.1243	3.3333	1200	0.4975	0.8513	0.8467
0.1243	3.6111	1300	0.6005	0.8358	0.8320
0.1243	3.8889	1400	0.6199	0.8226	0.8194
0.0931	4.1667	1500	0.5513	0.8464	0.8419
0.0931	4.4444	1600	0.5827	0.8409	0.8369
0.0931	4.7222	1700	0.6108	0.8374	0.8336
0.0672	5.0	1800	0.5761	0.8433	0.8392

Framework versions

• Transformers 5.3.0
• Pytorch 2.9.0+cu126
• Datasets 4.0.0
• Tokenizers 0.22.2