ModernBERT-large-classifier

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: 0.3071
• Accuracy: 0.9001
• F1: 0.8191

Training procedure

Training hyperparameters

The following hyperparameters were used during training:

• learning_rate: 2e-05
• train_batch_size: 32
• eval_batch_size: 32
• seed: 42
• 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: 3

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	F1
No log	0	0	0.7754	0.4796	0.5011
No log	0.6006	188	0.3604	0.8633	0.7299
No log	1.2013	376	0.2724	0.8897	0.7844
0.2959	1.8019	564	0.2549	0.9073	0.8343
0.2959	2.4026	752	0.3005	0.9033	0.8274

Framework versions

• Transformers 4.53.3
• Pytorch 2.6.0+cu124
• Datasets 4.0.0
• Tokenizers 0.21.2