ymoslem/ModernBERT-base-AIME-1983-2023-instruct-qe-classifier-binary-10ep-lr5e-05

Query Quality Estimation - Binary Classification

This model is a fine-tuned version of answerdotai/ModernBERT-base on the ymoslem/AIME-1983-2023-instruct dataset. It achieves the following results on the evaluation set:

• Loss: 0.8385
• Accuracy: 0.9
• F1 Macro: 0.8769
• F1 Weighted: 0.8982
• Precision: 0.8920
• Recall: 0.8651

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: 64
• eval_batch_size: 64
• 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: linear
• num_epochs: 10

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	F1 Macro	F1 Weighted	Precision	Recall
0.6792	1.0	29	0.4925	0.6333	0.6329	0.6378	0.725	0.7381
0.6136	2.0	58	0.4563	0.8333	0.7948	0.8304	0.8068	0.7857
0.439	3.0	87	0.3382	0.8	0.7885	0.8082	0.7879	0.8413
0.3324	4.0	116	0.2936	0.7833	0.7727	0.7924	0.7770	0.8294
0.2416	5.0	145	0.3051	0.8	0.7885	0.8082	0.7879	0.8413
0.2002	6.0	174	0.3490	0.8667	0.8535	0.8711	0.8403	0.8889
0.0714	7.0	203	0.5923	0.9167	0.9050	0.9183	0.8919	0.9246
0.0231	8.0	232	0.5090	0.8833	0.8703	0.8868	0.8561	0.9008
0.0057	9.0	261	0.6692	0.9	0.8845	0.9014	0.875	0.8968
0.0017	10.0	290	0.8385	0.9	0.8769	0.8982	0.8920	0.8651

Framework versions

• Transformers 4.57.0
• Pytorch 2.8.0+cu128
• Datasets 4.1.1
• Tokenizers 0.22.1