Modelpublic

Mithilss/gte-reranker-modernbert-base-finetune

sourceHugging Faceapache-2.0updated 1y agoView on Hugging Face

0likes4downloads

Model Card

gte-reranker-modernbert-base-finetune

This model is a fine-tuned version of Alibaba-NLP/gte-reranker-modernbert-base on an unknown dataset. It achieves the following results on the evaluation set:

—Loss: 0.4467
—Spearman: 0.3982

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: 6e-05
—trainbatchsize: 8
—evalbatchsize: 8
—seed: 42
—optimizer: Use OptimizerNames.ADAMW8BIT with betas=(0.9,0.999) and epsilon=1e-08 and optimizerargs=No additional optimizer arguments
—lrschedulertype: linear
—num_epochs: 3

Training results

Training Loss	Epoch	Step	Validation Loss	Spearman
No log	0	0	1.3782	-0.0348
0.2671	0.2458	1000	0.4733	0.0935
0.8379	0.4916	2000	0.4701	0.1638
0.772	0.7375	3000	0.4644	0.2124
0.2628	0.9833	4000	0.4654	0.2498
0.5863	1.2291	5000	0.4578	0.3172
0.6618	1.4749	6000	0.4576	0.3534
0.2003	1.7207	7000	0.4499	0.3697
0.5668	1.9666	8000	0.4455	0.3845
0.4551	2.2124	9000	0.4460	0.3914
0.5189	2.4582	10000	0.4461	0.3964
0.2746	2.7040	11000	0.4476	0.3980
0.6243	2.9499	12000	0.4467	0.3982

Framework versions

—Transformers 4.56.2
—Pytorch 2.8.0+cu128
—Datasets 3.6.0
—Tokenizers 0.22.1