Modelpublic

abdiharyadi/deberta-v3-large-ft-icar-a-v0.11

sourceHugging Facemitupdated 1y agoView on Hugging Face

0likes5downloads

Model Card

deberta-v3-large-ft-icar-a-v0.11

This model is a fine-tuned version of microsoft/deberta-v3-large on an unknown dataset. It achieves the following results on the evaluation set:

—Loss: 0.8879
—Accuracy: 0.9234
—Precision: 0.9164
—Recall: 0.9276
—F1: 0.9192

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: 3e-06
—trainbatchsize: 1
—evalbatchsize: 1
—seed: 42
—gradientaccumulationsteps: 3
—totaltrainbatch_size: 3
—optimizer: Use OptimizerNames.ADAMWTORCH with betas=(0.9,0.999) and epsilon=1e-08 and optimizerargs=No additional optimizer arguments
—lrschedulertype: linear
—num_epochs: 50

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	Precision	Recall	F1
2.3159	1.0	871	0.7036	0.8668	0.8807	0.8526	0.8620
1.7723	2.0	1742	0.8122	0.8897	0.9045	0.8675	0.8799
1.3618	3.0	2613	0.5902	0.9142	0.9024	0.9075	0.9046
1.1381	4.0	3484	0.7065	0.9188	0.9186	0.9003	0.9073
0.8802	5.0	4355	0.6828	0.9142	0.9068	0.8988	0.9019
0.6756	6.0	5226	0.7222	0.9219	0.9184	0.9107	0.9139
0.5417	7.0	6097	0.8736	0.9142	0.9209	0.8913	0.9032
0.5264	8.0	6968	0.8819	0.9035	0.8958	0.8927	0.8933
0.3627	9.0	7839	0.8570	0.9112	0.9013	0.9062	0.9037
0.3268	10.0	8710	0.8700	0.9204	0.9118	0.9161	0.9129
0.2345	11.0	9581	0.8820	0.9173	0.9090	0.9128	0.9095
0.1973	12.0	10452	0.9124	0.9219	0.9193	0.9088	0.9134
0.1236	13.0	11323	0.8879	0.9234	0.9164	0.9276	0.9192
0.1205	14.0	12194	1.0422	0.9066	0.9001	0.8869	0.8925
0.1324	15.0	13065	0.9978	0.9158	0.9070	0.9065	0.9063
0.0975	16.0	13936	1.0231	0.9142	0.9124	0.8972	0.9035
0.0793	17.0	14807	1.0438	0.9112	0.9077	0.8898	0.8965
0.1142	18.0	15678	1.0160	0.9142	0.9071	0.9170	0.9094

Framework versions

—Transformers 4.52.4
—Pytorch 2.6.0+cu124
—Datasets 4.0.0
—Tokenizers 0.21.2