vijil-bias-detector-v3b

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

• Loss: 0.2030
• Accuracy: 0.9194
• F1: 0.9200
• Precision: 0.8977
• Recall: 0.9434

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: 3
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	F1	Precision	Recall
0.2831	0.6242	500	0.3003	0.8776	0.8853	0.8196	0.9625
0.2473	1.2484	1000	0.2369	0.9007	0.8993	0.8953	0.9033
0.2137	1.8727	1500	0.2166	0.9060	0.9052	0.8964	0.9141
0.1733	2.4969	2000	0.2127	0.9191	0.9194	0.9001	0.9396
0.1660	3.0	2403	0.2030	0.9194	0.9200	0.8977	0.9434

Framework versions

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