results

This model is a fine-tuned version of google/electra-small-discriminator on the None dataset. It achieves the following results on the evaluation set:

• Loss: 0.1846
• Accuracy: 0.9573
• Precision: 0.9330
• Recall: 0.9766
• F1 Score: 0.9543
• True Positives: 376
• False Positives: 27
• True Negatives: 431
• False Negatives: 9

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: 16
• eval_batch_size: 16
• seed: 42
• optimizer: Use OptimizerNames.ADAMW_TORCH with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: linear
• num_epochs: 10

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	Precision	Recall	F1 Score	True Positives	False Positives	True Negatives	False Negatives
0.2007	1.0	246	0.1766	0.9514	0.9681	0.9336	0.9505	394	13	408	28
0.1446	2.0	492	0.1496	0.9609	0.9576	0.9645	0.9610	407	18	403	15
0.0992	3.0	738	0.1560	0.9597	0.9663	0.9526	0.9594	402	14	407	20
0.1022	4.0	984	0.1629	0.9609	0.9620	0.9597	0.9609	405	16	405	17
0.0876	5.0	1230	0.1472	0.9620	0.9643	0.9597	0.9620	405	15	406	17
0.0878	6.0	1476	0.1610	0.9609	0.9431	0.9810	0.9617	414	25	396	8
0.0779	7.0	1722	0.1644	0.9609	0.9555	0.9668	0.9611	408	19	402	14
0.0675	8.0	1968	0.1613	0.9644	0.9537	0.9763	0.9649	412	20	401	10
0.0816	9.0	2214	0.1617	0.9656	0.9538	0.9787	0.9661	413	20	401	9
0.0566	10.0	2460	0.1612	0.9656	0.9538	0.9787	0.9661	413	20	401	9

Framework versions

• Transformers 4.49.0
• Pytorch 2.6.0+cu124
• Datasets 3.4.1
• Tokenizers 0.21.1