PhoBert_content_48K

This model is a fine-tuned version of vinai/phobert-base-v2 on the None dataset. It achieves the following results on the evaluation set:

• Loss: 0.3045
• Accuracy: 0.9482
• F1: 0.9214

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: 64
• eval_batch_size: 64
• seed: 42
• 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: 10

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	F1
No log	0.3328	200	0.1642	0.9358	0.9006
No log	0.6656	400	0.1668	0.9388	0.9042
0.192	0.9983	600	0.1516	0.9437	0.9128
0.192	1.3311	800	0.1352	0.9485	0.9225
0.192	1.6639	1000	0.1412	0.9483	0.9214
0.126	1.9967	1200	0.1296	0.9520	0.9302
0.126	2.3295	1400	0.1692	0.9437	0.9121
0.126	2.6622	1600	0.1534	0.9527	0.9306
0.0963	2.9950	1800	0.1658	0.9503	0.9247
0.0963	3.3278	2000	0.1755	0.9479	0.9208
0.0963	3.6606	2200	0.1677	0.9485	0.9221
0.0733	3.9933	2400	0.1712	0.9504	0.9249
0.0733	4.3261	2600	0.1811	0.9502	0.9255
0.0733	4.6589	2800	0.1930	0.9496	0.9255
0.0536	4.9917	3000	0.1870	0.9488	0.9222
0.0536	5.3245	3200	0.2496	0.9444	0.9135
0.0536	5.6572	3400	0.2162	0.9501	0.9236
0.0425	5.9900	3600	0.2401	0.9463	0.9160
0.0425	6.3228	3800	0.2399	0.9504	0.9259
0.0425	6.6556	4000	0.2495	0.9490	0.9228
0.032	6.9884	4200	0.2818	0.9481	0.9198
0.032	7.3211	4400	0.2611	0.9485	0.9221
0.032	7.6539	4600	0.2723	0.9496	0.9243
0.0219	7.9867	4800	0.2711	0.9483	0.9221
0.0219	8.3195	5000	0.2876	0.9485	0.9233
0.0219	8.6522	5200	0.3138	0.9482	0.9210
0.0185	8.9850	5400	0.2994	0.9482	0.9205
0.0185	9.3178	5600	0.2899	0.9486	0.9224
0.0185	9.6506	5800	0.3040	0.9476	0.9205
0.013	9.9834	6000	0.3045	0.9482	0.9214

Framework versions

• Transformers 4.53.0
• Pytorch 2.7.1+cu126
• Datasets 3.6.0
• Tokenizers 0.21.0