UNEDMediaBiasTeam_at_SemEval23_Task3_Subtask3_PRE_BABE_dataset

This model is a fine-tuned version of distilbert-base-uncased-finetuned-sst-2-english on the None dataset. It achieves the following results on the evaluation set:

• Loss: 1.2188
• F1: 0.5660

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: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 30
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	F1
0.5684	1.0	24	0.8998	0.4355
0.3	2.0	48	0.9073	0.4625
0.3634	3.0	72	0.8815	0.4868
0.2344	4.0	96	0.9457	0.4848
0.1712	5.0	120	0.9737	0.4945
0.148	6.0	144	1.0416	0.4896
0.0662	7.0	168	1.1345	0.4838
0.046	8.0	192	1.0935	0.5353
0.0398	9.0	216	1.1288	0.5376
0.0563	10.0	240	1.2188	0.5660
0.0449	11.0	264	1.2390	0.5160
0.0472	12.0	288	1.3779	0.5069
0.0122	13.0	312	1.4218	0.5442
0.0037	14.0	336	1.4859	0.5432
0.0557	15.0	360	1.5124	0.5510
0.0038	16.0	384	1.5364	0.5542
0.0043	17.0	408	1.5484	0.5589
0.0022	18.0	432	1.6063	0.5554
0.0044	19.0	456	1.6013	0.5268
0.0023	20.0	480	1.6161	0.4802
0.002	21.0	504	1.6622	0.4783
0.0016	22.0	528	1.6737	0.4812
0.002	23.0	552	1.6776	0.5250
0.0019	24.0	576	1.7027	0.4800
0.0015	25.0	600	1.6897	0.5211
0.0018	26.0	624	1.6982	0.5211
0.0015	27.0	648	1.7174	0.4781
0.0019	28.0	672	1.7269	0.4781
0.0016	29.0	696	1.7323	0.5133
0.0304	30.0	720	1.7265	0.5172

Framework versions

• Transformers 4.26.0
• Pytorch 1.13.1
• Datasets 2.9.0
• Tokenizers 0.13.2