modernbert-3pair-adv-3label-v2

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

• Loss: 0.3307
• Accuracy: 0.9216
• F1: 0.9214
• Precision: 0.9215
• Recall: 0.9216
• F1 Class 0: 0.9227
• Precision Class 0: 0.9204
• Recall Class 0: 0.9250
• F1 Class 1: 0.9399
• Precision Class 1: 0.9285
• Recall Class 1: 0.9516
• F1 Class 2: 0.9019
• Precision Class 2: 0.9156
• Recall Class 2: 0.8886

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: 8e-06
• train_batch_size: 32
• eval_batch_size: 32
• seed: 42
• distributed_type: multi-GPU
• num_devices: 6
• total_train_batch_size: 192
• total_eval_batch_size: 192
• 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
• lr_scheduler_warmup_ratio: 0.1
• num_epochs: 1
• label_smoothing_factor: 0.05

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	F1	Precision	Recall	F1 Class 0	Precision Class 0	Recall Class 0	F1 Class 1	Precision Class 1	Recall Class 1	F1 Class 2	Precision Class 2	Recall Class 2
1.9049	1.0	13603	0.3307	0.9216	0.9214	0.9215	0.9216	0.9227	0.9204	0.9250	0.9399	0.9285	0.9516	0.9019	0.9156	0.8886

Framework versions

• Transformers 4.55.0
• Pytorch 2.8.0+cu128
• Datasets 4.0.0
• Tokenizers 0.21.4