scenario-TCR_data-cl-cardiff_cl_only2

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: 4.5723
• Accuracy: 0.5077
• F1: 0.5075

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: 1244
• 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	1.0569	0.4907	0.4770
0.9819	2.17	500	1.1934	0.4699	0.4678
0.9819	3.26	750	1.2411	0.5154	0.5095
0.6473	4.35	1000	1.5514	0.5069	0.5070
0.6473	5.43	1250	1.7216	0.5	0.4995
0.389	6.52	1500	2.1097	0.4807	0.4781
0.389	7.61	1750	2.1021	0.4769	0.4742
0.2296	8.7	2000	2.2799	0.4900	0.4912
0.2296	9.78	2250	2.7726	0.5008	0.5024
0.1512	10.87	2500	2.9023	0.4923	0.4924
0.1512	11.96	2750	2.9147	0.5062	0.5051
0.1123	13.04	3000	2.8971	0.5201	0.5203
0.1123	14.13	3250	3.4501	0.4961	0.4977
0.0787	15.22	3500	3.3121	0.4877	0.4865
0.0787	16.3	3750	3.2467	0.4961	0.4982
0.0574	17.39	4000	3.7087	0.4684	0.4655
0.0574	18.48	4250	3.8077	0.5046	0.5069
0.0423	19.57	4500	3.9974	0.5	0.5003
0.0423	20.65	4750	4.0173	0.5023	0.5015
0.0245	21.74	5000	4.0356	0.5	0.5004
0.0245	22.83	5250	4.2308	0.4992	0.4996
0.0177	23.91	5500	4.3988	0.4900	0.4885
0.0177	25.0	5750	4.2642	0.5085	0.5093
0.0113	26.09	6000	4.3893	0.5062	0.5053
0.0113	27.17	6250	4.5465	0.4977	0.4979
0.0076	28.26	6500	4.5135	0.5093	0.5098
0.0076	29.35	6750	4.5723	0.5077	0.5075

Framework versions

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