mental-roberta-base-finetuned-depression

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

• Loss: 0.6567
• Precision: 0.8863
• Recall: 0.9168
• F1: 0.8996
• Accuracy: 0.9115

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: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 20

Training results

Training Loss	Epoch	Step	Validation Loss	Precision	Recall	F1	Accuracy
No log	1.0	469	0.3852	0.7878	0.8253	0.7958	0.8667
0.5249	2.0	938	0.4720	0.8778	0.8722	0.8662	0.8913
0.2598	3.0	1407	0.5459	0.8975	0.8791	0.8865	0.8977
0.1624	4.0	1876	0.5022	0.9004	0.8979	0.8976	0.9072
0.1036	5.0	2345	0.6257	0.8910	0.8968	0.8931	0.9009
0.0668	6.0	2814	0.6531	0.9145	0.8927	0.9026	0.9104
0.0539	7.0	3283	0.6209	0.8552	0.9115	0.8802	0.8945
0.057	8.0	3752	0.6567	0.8863	0.9168	0.8996	0.9115
0.0523	9.0	4221	0.7184	0.9067	0.8984	0.8993	0.9083
0.0354	10.0	4690	0.7112	0.8874	0.9014	0.8914	0.9072
0.0268	11.0	5159	0.7168	0.8996	0.9012	0.8979	0.9083
0.0297	12.0	5628	0.7499	0.8667	0.9096	0.8847	0.9030
0.0242	13.0	6097	0.7554	0.8946	0.9014	0.8955	0.9072
0.0238	14.0	6566	0.7990	0.8909	0.9014	0.8934	0.9072
0.0178	15.0	7035	0.8298	0.8965	0.8933	0.8925	0.9051
0.0226	16.0	7504	0.8428	0.9099	0.8890	0.8973	0.9062
0.0226	17.0	7973	0.8490	0.8742	0.8983	0.8816	0.9041
0.0183	18.0	8442	0.8148	0.8940	0.8965	0.8930	0.9072
0.0188	19.0	8911	0.8146	0.8927	0.8960	0.8921	0.9062
0.015	20.0	9380	0.8216	0.8927	0.8960	0.8921	0.9062

Framework versions

• Transformers 4.41.2
• Pytorch 2.3.0+cu121
• Datasets 2.20.0
• Tokenizers 0.19.1