3adapter_backbone_100k_v3

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:

• Acc Content: 0.9710
• F1 Content: 0.9664
• Loss: 0.0614

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
• gradient_accumulation_steps: 2
• total_train_batch_size: 128
• 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: 10

Training results

Training Loss	Epoch	Step	Acc Content	F1 Content	Validation Loss
0.0243	1.0	529	0.9693	0.9647	0.0202
0.0163	2.0	1058	0.9722	0.9677	0.0199
0.012	3.0	1587	0.9683	0.9637	0.0211
0.0097	4.0	2116	0.9710	0.9666	0.0279
0.0063	5.0	2645	0.9700	0.9653	0.0375
0.0052	6.0	3174	0.9696	0.9651	0.0413
0.0034	7.0	3703	0.9713	0.9668	0.0494
0.0026	8.0	4232	0.9706	0.9660	0.0540
0.0017	9.0	4761	0.9708	0.9662	0.0576
0.0012	10.0	5290	0.9710	0.9664	0.0614

Framework versions

• Transformers 4.57.1
• Pytorch 2.6.0+cu124
• Datasets 4.4.1
• Tokenizers 0.22.1