CNEC2_0_Supertypes_xlm-roberta-large

This model is a fine-tuned version of FacebookAI/xlm-roberta-large on the cnec dataset. It achieves the following results on the evaluation set:

• Loss: 0.2155
• Precision: 0.8565
• Recall: 0.8971
• F1: 0.8763
• Accuracy: 0.9709

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: 15

Training results

Training Loss	Epoch	Step	Validation Loss	Precision	Recall	F1	Accuracy
0.4393	1.0	900	0.1671	0.7756	0.8195	0.7969	0.9590
0.1716	2.0	1800	0.1409	0.8155	0.8583	0.8364	0.9662
0.1326	3.0	2700	0.1288	0.8203	0.8748	0.8467	0.9687
0.1027	4.0	3600	0.1408	0.8290	0.8732	0.8505	0.9683
0.0891	5.0	4500	0.1447	0.8485	0.9000	0.8735	0.9725
0.0715	6.0	5400	0.1393	0.8561	0.8868	0.8712	0.9713
0.0644	7.0	6300	0.1586	0.8517	0.8918	0.8713	0.9702
0.0535	8.0	7200	0.1526	0.8481	0.8810	0.8643	0.9696
0.0492	9.0	8100	0.1795	0.8529	0.8984	0.8751	0.9702
0.0391	10.0	9000	0.1903	0.8536	0.8938	0.8733	0.9693
0.0323	11.0	9900	0.1885	0.8615	0.9046	0.8825	0.9724
0.0274	12.0	10800	0.2099	0.8585	0.9025	0.8800	0.9696
0.0237	13.0	11700	0.1944	0.8624	0.9009	0.8812	0.9720
0.0245	14.0	12600	0.2129	0.8618	0.8967	0.8789	0.9711
0.0206	15.0	13500	0.2155	0.8565	0.8971	0.8763	0.9709

Framework versions

• Transformers 4.36.2
• Pytorch 2.1.2+cu121
• Datasets 2.16.1
• Tokenizers 0.15.0