MARBERT

This model is a fine-tuned version of UBC-NLP/MARBERT on the None dataset. It achieves the following results on the evaluation set:

• Loss: 1.0270
• Precision: 0.4848
• Recall: 0.6957
• F1: 0.5714
• Accuracy: 0.6772

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: 8
• eval_batch_size: 8
• 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	Precision	Recall	F1	Accuracy
No log	1.0	5	0.8519	0.0	0.0	0.0	0.625
No log	2.0	10	0.8108	0.0	0.0	0.0	0.6806
No log	3.0	15	0.7524	0.0	0.0	0.0	0.7361
No log	4.0	20	0.6927	0.2222	0.2857	0.25	0.7083
No log	5.0	25	0.6995	0.0	0.0	0.0	0.7222
No log	6.0	30	0.8241	0.1818	0.2857	0.2222	0.625
No log	7.0	35	0.6811	0.2222	0.2857	0.25	0.75
No log	8.0	40	0.6957	0.25	0.2857	0.2667	0.7639
No log	9.0	45	0.7392	0.2222	0.2857	0.25	0.6944
No log	10.0	50	0.7612	0.2222	0.2857	0.25	0.6944

Framework versions

• Transformers 4.48.1
• Pytorch 2.5.1+cu124
• Datasets 3.2.0
• Tokenizers 0.21.0