phobert-v2_v2

This model is a fine-tuned version of vinai/phobert-base-v2 on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 0.3130
• Accuracy: 0.9507
• Precision Macro: 0.8904
• Recall Macro: 0.8541
• F1 Macro: 0.8704
• F1 Weighted: 0.9497

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: 3e-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: linear
• 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
0.4636	1.0	90	0.2180	0.9419	0.9099	0.7652	0.8049	0.9359
0.1882	2.0	180	0.1916	0.9419	0.8351	0.8649	0.8485	0.9433
0.1453	3.0	270	0.1898	0.9488	0.8743	0.8402	0.8555	0.9476
0.1175	4.0	360	0.1932	0.9526	0.9141	0.8267	0.8597	0.9500
0.0856	5.0	450	0.2092	0.9514	0.8708	0.8711	0.8709	0.9514
0.0826	6.0	540	0.2221	0.9526	0.9063	0.8516	0.8748	0.9512
0.0675	7.0	630	0.2342	0.9438	0.8419	0.8696	0.8545	0.9450
0.0618	8.0	720	0.2402	0.9469	0.8890	0.8430	0.8630	0.9456
0.0426	9.0	810	0.2503	0.9507	0.8797	0.8543	0.8660	0.9499
0.038	10.0	900	0.2786	0.9514	0.8999	0.8467	0.8692	0.9499
0.039	11.0	990	0.2795	0.9463	0.8628	0.8554	0.8589	0.9460
0.0263	12.0	1080	0.2817	0.9488	0.8733	0.8571	0.8648	0.9483
0.0209	13.0	1170	0.2840	0.9495	0.8802	0.8576	0.8681	0.9488
0.0221	14.0	1260	0.2769	0.9526	0.8904	0.8639	0.8761	0.9519
0.0172	15.0	1350	0.2861	0.9514	0.8985	0.8546	0.8739	0.9502
0.0159	16.0	1440	0.3031	0.9482	0.8850	0.8523	0.8671	0.9472
0.0124	17.0	1530	0.3119	0.9501	0.8792	0.8538	0.8655	0.9493
0.0107	18.0	1620	0.3173	0.9476	0.8948	0.8476	0.8681	0.9463
0.0106	19.0	1710	0.3094	0.9545	0.8979	0.8611	0.8776	0.9535
0.0108	20.0	1800	0.3130	0.9507	0.8904	0.8541	0.8704	0.9497

Framework versions

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