Savoxism/ModernBERT-base_Ver1

ver_1

This model is a fine-tuned version of answerdotai/ModernBERT-base on the None dataset. It achieves the following results on the evaluation set:

• Loss: 0.4152
• Map@3: 0.9329

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: 5e-05
• train_batch_size: 32
• eval_batch_size: 64
• 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
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Map@3
0.7525	0.2179	200	0.7113	0.8590
0.5624	0.4357	400	0.5364	0.8959
0.5259	0.6536	600	0.5016	0.9052
0.4445	0.8715	800	0.4248	0.9185
0.336	1.0893	1000	0.4341	0.9196
0.355	1.3072	1200	0.4060	0.9233
0.3176	1.5251	1400	0.3799	0.9274
0.3067	1.7429	1600	0.3695	0.9322
0.3175	1.9608	1800	0.3597	0.9331
0.1646	2.1786	2000	0.4279	0.9330
0.1564	2.3965	2200	0.4206	0.9311
0.1495	2.6144	2400	0.4173	0.9331
0.1791	2.8322	2600	0.4152	0.9329

Framework versions

• Transformers 4.52.4
• Pytorch 2.6.0+cu124
• Datasets 3.6.0
• Tokenizers 0.21.2