Hashuz/QT_MentalQA

results

This model is a fine-tuned version of UBC-NLP/MARBERTv2 on the None dataset. It achieves the following results on the evaluation set:

• Loss: 0.3179
• F1-micro: 0.7925
• F1-macro: 0.7727
• Jaccard: 0.7206

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: 8
• eval_batch_size: 8
• seed: 42
• 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: 15

Training results

Training Loss	Epoch	Step	Validation Loss	F1-micro	F1-macro	Jaccard
No log	1.0	39	0.4793	0.5859	0.4521	0.4773
No log	2.0	78	0.4556	0.5737	0.4510	0.4754
0.4791	3.0	117	0.4079	0.6209	0.4892	0.5413
0.4791	4.0	156	0.3969	0.5785	0.4580	0.5008
0.4791	5.0	195	0.3711	0.6526	0.5956	0.5738
0.3661	6.0	234	0.3866	0.6246	0.5735	0.5357
0.3661	7.0	273	0.3400	0.7123	0.6774	0.6405
0.2748	8.0	312	0.3415	0.7163	0.6823	0.6365
0.2748	9.0	351	0.3296	0.75	0.7204	0.6714
0.2748	10.0	390	0.3162	0.7735	0.7438	0.6984
0.2162	11.0	429	0.3174	0.7859	0.7613	0.7111
0.2162	12.0	468	0.3056	0.7847	0.7656	0.7222
0.1889	13.0	507	0.3217	0.7880	0.7681	0.7159
0.1889	14.0	546	0.3209	0.7847	0.7649	0.7111
0.1889	15.0	585	0.3179	0.7925	0.7727	0.7206

Framework versions

• Transformers 4.52.4
• Pytorch 2.6.0+cu124
• Datasets 2.14.4
• Tokenizers 0.21.1