Modelpublic

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

sourceHugging Facemitupdated 10mo agoView on Hugging Face

0likes43downloads

Model Card

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

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: 0.8877
—Accuracy: 0.9149
—Precision: 0.8989
—Recall: 0.8870
—F1: 0.8925

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
2.2451	1.0	704	0.7087	0.8450	0.8182	0.7916	0.8022
1.7218	2.0	1408	0.7015	0.8771	0.8588	0.8321	0.8435
1.339	3.0	2112	0.8289	0.8790	0.8768	0.8236	0.8425
1.0953	4.0	2816	0.8203	0.8922	0.8892	0.8474	0.8619
0.8391	5.0	3520	0.7416	0.9055	0.8888	0.8803	0.8842
0.7217	6.0	4224	0.7420	0.9074	0.8879	0.8796	0.8835
0.6325	7.0	4928	0.7878	0.8960	0.9077	0.8313	0.8557
0.4921	8.0	5632	0.8229	0.8979	0.8749	0.8653	0.8698
0.4206	9.0	6336	0.9300	0.8866	0.8583	0.8624	0.8603
0.3445	10.0	7040	0.8305	0.9055	0.8881	0.8733	0.8799
0.2187	11.0	7744	1.0823	0.8809	0.8565	0.8357	0.8447
0.2072	12.0	8448	0.8940	0.9055	0.8939	0.8599	0.8723
0.1579	13.0	9152	0.9451	0.8979	0.8706	0.8640	0.8670
0.209	14.0	9856	0.8635	0.9036	0.8934	0.8643	0.8767
0.1432	15.0	10560	0.9687	0.9017	0.8992	0.8497	0.8678
0.0898	16.0	11264	0.8877	0.9149	0.8989	0.8870	0.8925
0.0516	17.0	11968	1.0159	0.9055	0.8930	0.8594	0.8723
0.0682	18.0	12672	1.1908	0.8866	0.8519	0.8718	0.8586
0.182	19.0	13376	0.9440	0.9112	0.8895	0.8935	0.8914
0.0817	20.0	14080	1.0778	0.8979	0.8846	0.8653	0.8739
0.1148	21.0	14784	1.0861	0.9055	0.8860	0.8746	0.8798
0.1034	22.0	15488	1.1888	0.8847	0.8585	0.8422	0.8485
0.108	23.0	16192	0.9938	0.9074	0.8900	0.8894	0.8895
0.032	24.0	16896	1.2295	0.8960	0.8803	0.8585	0.8677
0.0975	25.0	17600	1.1770	0.8998	0.8912	0.8679	0.8782
0.0226	26.0	18304	1.2022	0.8979	0.8869	0.8666	0.8757
0.0527	27.0	19008	1.0817	0.9093	0.9004	0.8760	0.8868
0.0174	28.0	19712	1.0276	0.9130	0.8942	0.8911	0.8925
0.061	29.0	20416	1.0123	0.9149	0.9030	0.8814	0.8904
0.0496	30.0	21120	1.0771	0.9093	0.8990	0.8716	0.8829
0.0436	31.0	21824	1.1075	0.9036	0.8809	0.8885	0.8845

Framework versions

—Transformers 4.52.4
—Pytorch 2.6.0+cu124
—Datasets 4.4.1
—Tokenizers 0.21.2