Modelpublic

ymoslem/ModernBERT-base-TeleQnA-router-qe-classifier-binary-10ep-lr2e-05-qwen4b-5runs-5evals

sourceHugging Faceapache-2.0updated 4mo agoView on Hugging Face

0likes6downloads

Model Card

Query Quality Estimation - Binary Classification

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

—Loss: 3.1982
—Accuracy: 0.665
—F1 Macro: 0.5873
—F1 Weighted: 0.6479
—Precision: 0.6048
—Recall: 0.5852

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: 2e-05
—trainbatchsize: 64
—evalbatchsize: 64
—seed: 42
—optimizer: Use OptimizerNames.ADAMWTORCHFUSED with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
—lrschedulertype: linear
—num_epochs: 10

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	F1 Macro	F1 Weighted	Precision	Recall
0.3857	1.0	704	1.0551	0.6658	0.5596	0.6328	0.5971	0.5643
0.1486	2.0	1408	1.3319	0.6342	0.5883	0.6348	0.5880	0.5887
0.0954	3.0	2112	1.8186	0.6512	0.5644	0.6302	0.5829	0.5648
0.0279	4.0	2816	1.8675	0.6262	0.6014	0.6350	0.6017	0.6119
0.0155	5.0	3520	2.6794	0.6638	0.5781	0.6425	0.6004	0.5774
0.0123	6.0	4224	2.7620	0.6458	0.5795	0.6360	0.5867	0.5776
0.0083	7.0	4928	2.7599	0.6524	0.5930	0.6456	0.5982	0.5908
0.002	8.0	5632	3.1982	0.665	0.5873	0.6479	0.6048	0.5852

Framework versions

—Transformers 4.57.3
—Pytorch 2.8.0+cu128
—Datasets 4.1.1
—Tokenizers 0.22.1