vijil-bias-detector-v3

This model is a fine-tuned version of on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 0.1666
• Accuracy: 0.9157
• F1: 0.9160
• Precision: 0.8965
• Recall: 0.9364

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: 32
• eval_batch_size: 64
• seed: 42
• optimizer: Use OptimizerNames.ADAMW_TORCH_FUSED with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: linear
• lr_scheduler_warmup_steps: 0.1
• num_epochs: 5
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	F1	Precision	Recall
0.6340	0.2497	200	0.5716	0.7364	0.7100	0.7720	0.6571
0.4299	0.4994	400	0.4346	0.7464	0.7019	0.8299	0.6081
0.3864	0.7491	600	0.3375	0.7864	0.7992	0.7421	0.8658
0.3235	0.9988	800	0.3435	0.7701	0.7260	0.8752	0.6202
0.3131	1.2484	1000	0.2941	0.8235	0.8265	0.7988	0.8562
0.2746	1.4981	1200	0.3141	0.8032	0.7764	0.8780	0.6959
0.1990	1.7478	1400	0.2109	0.8894	0.8913	0.8612	0.9237
0.1862	1.9975	1600	0.1868	0.8994	0.9021	0.8642	0.9434
0.1882	2.2472	1800	0.1825	0.9069	0.9065	0.8947	0.9186
0.1755	2.4969	2000	0.1748	0.9101	0.9118	0.8790	0.9472
0.1558	2.7466	2200	0.1721	0.9129	0.9131	0.8949	0.9319
0.1583	2.9963	2400	0.1773	0.9041	0.9069	0.8667	0.9510
0.1349	3.2459	2600	0.1795	0.9082	0.9061	0.9101	0.9020
0.1323	3.4956	2800	0.1652	0.9132	0.9121	0.9069	0.9173
0.1352	3.7453	3000	0.1664	0.9151	0.9154	0.8954	0.9364
0.1388	3.9950	3200	0.1623	0.9185	0.9191	0.8961	0.9434
0.1134	4.2447	3400	0.1713	0.9138	0.9130	0.905	0.9211
0.1056	4.4944	3600	0.1678	0.9182	0.9184	0.8999	0.9377
0.1123	4.7441	3800	0.1665	0.9157	0.9156	0.8999	0.9319
0.0985	4.9938	4000	0.1666	0.9157	0.9160	0.8965	0.9364
0.0985	5.0	4005	0.1666	0.9157	0.9160	0.8965	0.9364

Framework versions

• Transformers 5.5.0
• Pytorch 2.11.0+cu130
• Datasets 4.8.4
• Tokenizers 0.22.2