deberta_sent4455

This model is a fine-tuned version of kisti/korscideberta on the None dataset. It achieves the following results on the evaluation set:

• Loss: 0.3113
• Accuracy: 0.9066
• Precision: 0.8995
• Recall: 0.9066
• F1: 0.9011

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: 1.5e-05
• train_batch_size: 16
• eval_batch_size: 8
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_steps: 500
• num_epochs: 5
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	Precision	Recall	F1
0.32	0.3934	500	0.2873	0.8974	0.8870	0.8974	0.8878
0.284	0.7868	1000	0.2800	0.9022	0.8934	0.9022	0.8897
0.2797	1.1802	1500	0.2648	0.9037	0.8977	0.9037	0.8997
0.2759	1.5736	2000	0.2597	0.9107	0.9036	0.9107	0.9038
0.2746	1.9670	2500	0.2568	0.9100	0.9027	0.9100	0.9030
0.2523	2.3603	3000	0.3300	0.8793	0.8898	0.8793	0.8837
0.232	2.7537	3500	0.2753	0.9077	0.8999	0.9077	0.8996
0.217	3.1471	4000	0.2962	0.9096	0.9027	0.9096	0.9039
0.2265	3.5405	4500	0.2787	0.9107	0.9035	0.9107	0.9034
0.2094	3.9339	5000	0.2944	0.9085	0.9008	0.9085	0.8999
0.1933	4.3273	5500	0.3132	0.9066	0.8999	0.9066	0.9017
0.1935	4.7207	6000	0.3113	0.9066	0.8995	0.9066	0.9011

Framework versions

• Transformers 4.45.0
• Pytorch 2.6.0+cu124
• Datasets 2.14.4
• Tokenizers 0.20.3