ModernBERT-CLINC_au

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: 0.7736
• Accuracy: 0.9410
• F1 Macro: 0.9457

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: 4e-06
• train_batch_size: 24
• eval_batch_size: 24
• 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: cosine
• lr_scheduler_warmup_ratio: 0.06
• num_epochs: 4
• label_smoothing_factor: 0.05

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	F1 Macro
5.0084	0.3145	200	4.5732	0.0961	0.0746
3.1036	0.6289	400	1.9721	0.6426	0.6316
1.2603	0.9434	600	1.0996	0.8535	0.8589
0.794	1.2579	800	0.9419	0.8881	0.8939
0.7224	1.5723	1000	0.8649	0.9103	0.9146
0.6899	1.8868	1200	0.8257	0.9229	0.9277
0.6041	2.2013	1400	0.8044	0.9303	0.9353
0.5701	2.5157	1600	0.7925	0.9355	0.9400
0.5611	2.8302	1800	0.7831	0.9361	0.9407
0.5427	3.1447	2000	0.7813	0.9377	0.9425
0.5288	3.4591	2200	0.7731	0.9406	0.9451
0.5259	3.7736	2400	0.7736	0.9410	0.9457

Framework versions

• Transformers 4.56.1
• Pytorch 2.8.0+cu126
• Datasets 4.0.0
• Tokenizers 0.22.0