LILT-id

This model is a fine-tuned version of nielsr/lilt-xlm-roberta-base on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 0.4723
• Precision: 0.9132
• Recall: 0.8998
• F1: 0.9064
• Accuracy: 0.9467

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: 2
• eval_batch_size: 4
• 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	Precision	Recall	F1	Accuracy
No log	2.4390	200	0.2933	0.875	0.8557	0.8653	0.9208
No log	4.8780	400	0.3650	0.8934	0.8606	0.8767	0.9257
0.3014	7.3171	600	0.3822	0.9055	0.8900	0.8977	0.9386
0.3014	9.7561	800	0.4052	0.8980	0.8826	0.8903	0.9386
0.0365	12.1951	1000	0.4668	0.8966	0.8900	0.8933	0.9386
0.0365	14.6341	1200	0.4664	0.9123	0.8900	0.9010	0.9435
0.0365	17.0732	1400	0.4993	0.8978	0.8802	0.8889	0.9370
0.0091	19.5122	1600	0.4723	0.9132	0.8998	0.9064	0.9467
0.0091	21.9512	1800	0.4826	0.9089	0.9022	0.9055	0.9467
0.0004	24.3902	2000	0.4790	0.9086	0.8998	0.9042	0.9467
0.0004	26.8293	2200	0.4807	0.9086	0.8998	0.9042	0.9467
0.0004	29.2683	2400	0.4818	0.9086	0.8998	0.9042	0.9467

Framework versions

• Transformers 4.42.4
• Pytorch 2.3.1+cu121
• Datasets 2.20.0
• Tokenizers 0.19.1