vijil-bias-detector-v4

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

• Loss: 0.1877
• Accuracy: 0.9050
• F1: 0.9094
• Precision: 0.8983
• Recall: 0.9207

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
0.5963	0.3221	200	0.4896	0.7339	0.7245	0.7808	0.6757
0.3916	0.6441	400	0.3866	0.7681	0.7700	0.7915	0.7496
0.3415	0.9662	600	0.3335	0.7959	0.8076	0.7887	0.8274
0.3058	1.2882	800	0.3029	0.8172	0.8225	0.8270	0.8180
0.2417	1.6103	1000	0.2334	0.8720	0.8797	0.8564	0.9044
0.2137	1.9324	1200	0.2192	0.8853	0.8891	0.8901	0.8880
0.1833	2.2544	1400	0.2099	0.8921	0.9011	0.8574	0.9495
0.1805	2.5765	1600	0.2022	0.8937	0.8992	0.8831	0.9160
0.1733	2.8986	1800	0.2048	0.8869	0.8861	0.9255	0.8499
0.1473	3.2206	2000	0.1850	0.9058	0.9116	0.8868	0.9378
0.1271	3.5427	2200	0.1941	0.9066	0.9121	0.8892	0.9362
0.1159	3.8647	2400	0.1865	0.9070	0.9108	0.9042	0.9176
0.0996	4.1868	2600	0.1906	0.9034	0.9076	0.8986	0.9168
0.1017	4.5089	2800	0.1923	0.9022	0.9078	0.8866	0.9300
0.0952	4.8309	3000	0.1902	0.9046	0.9099	0.8905	0.9300
0.0945	5.0	3105	0.1877	0.9050	0.9094	0.8983	0.9207

Framework versions

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