scenario-NON-KD-SCR-COPY-CDF-CL-D2_data-cl-cardiff_cl_only_alpha

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

• Loss: 6.7853
• Accuracy: 0.3580
• F1: 0.3378

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: 32
• eval_batch_size: 32
• seed: 1123
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 30

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	F1
No log	1.09	250	1.1518	0.3495	0.3449
1.0509	2.17	500	1.5848	0.3665	0.3279
1.0509	3.26	750	2.0537	0.3673	0.3351
0.4971	4.35	1000	2.6254	0.3642	0.3389
0.4971	5.43	1250	3.3984	0.3495	0.3056
0.189	6.52	1500	3.9545	0.3588	0.3213
0.189	7.61	1750	4.3147	0.3634	0.3250
0.0832	8.7	2000	4.5326	0.3495	0.3223
0.0832	9.78	2250	4.8999	0.3627	0.3396
0.0407	10.87	2500	5.2749	0.3503	0.3354
0.0407	11.96	2750	5.2814	0.3634	0.3500
0.0279	13.04	3000	5.3923	0.3657	0.3502
0.0279	14.13	3250	5.7450	0.3565	0.3397
0.0153	15.22	3500	5.6113	0.3681	0.3582
0.0153	16.3	3750	5.1689	0.3704	0.3615
0.0145	17.39	4000	5.8264	0.3650	0.3579
0.0145	18.48	4250	5.6710	0.3650	0.3603
0.0092	19.57	4500	6.0070	0.3650	0.3547
0.0092	20.65	4750	6.2579	0.3519	0.3287
0.0034	21.74	5000	6.3540	0.3650	0.3550
0.0034	22.83	5250	6.4666	0.3619	0.3438
0.0031	23.91	5500	6.6982	0.3580	0.3286
0.0031	25.0	5750	6.6139	0.3657	0.3611
0.0016	26.09	6000	6.6320	0.3688	0.3632
0.0016	27.17	6250	6.7619	0.3673	0.3544
0.0015	28.26	6500	6.7368	0.3627	0.3426
0.0015	29.35	6750	6.7853	0.3580	0.3378

Framework versions

• Transformers 4.33.3
• Pytorch 2.1.1+cu121
• Datasets 2.14.5
• Tokenizers 0.13.3