ai_vs_human_detector_deberta_v3_robust

This model is a fine-tuned version of microsoft/deberta-v3-large on the None dataset. It achieves the following results on the evaluation set:

• Loss: 0.0228
• Accuracy: 0.9940
• F1: 0.9939

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: 1e-05
• train_batch_size: 8
• eval_batch_size: 8
• seed: 42
• gradient_accumulation_steps: 4
• total_train_batch_size: 32
• 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: 1
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	F1
0.1372	0.1082	500	0.0982	0.9631	0.9638
0.0362	0.2164	1000	0.0379	0.9891	0.9890
0.037	0.3246	1500	0.0409	0.9884	0.9884
0.0325	0.4328	2000	0.0287	0.9919	0.9918
0.0222	0.5410	2500	0.0262	0.9943	0.9943
0.0222	0.6492	3000	0.0209	0.9945	0.9944
0.0218	0.7575	3500	0.0494	0.9879	0.9879
0.019	0.8657	4000	0.0198	0.9950	0.9949
0.0226	0.9739	4500	0.0228	0.9940	0.9939

Framework versions

• Transformers 4.57.3
• Pytorch 2.7.1+cu118
• Datasets 4.4.1
• Tokenizers 0.22.1