SignBart-KArSL03-502

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

• Loss: 0.0091
• Accuracy: 0.9980
• Precision: 0.9985
• Recall: 0.9980

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: 0.0002
• train_batch_size: 256
• eval_batch_size: 256
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: cosine
• num_epochs: 1000

Training results

Training Loss	Epoch	Step	Accuracy	Validation Loss	Precision	Recall
6.6088	1.0	83	0.0357	5.6758	0.0164	0.0357
5.9921	2.0	166	0.1518	4.6968	0.1139	0.1518
5.2424	3.0	249	0.3689	3.8356	0.3420	0.3689
4.6038	4.0	332	0.5414	3.0803	0.5276	0.5414
3.9913	5.0	415	0.6835	2.3784	0.6970	0.6835
3.511	6.0	498	0.7773	1.8461	0.7960	0.7773
2.9875	7.0	581	0.8397	1.3705	0.8500	0.8397
2.5896	8.0	664	0.8814	1.0179	0.8902	0.8814
2.1975	9.0	747	0.9101	0.7636	0.9195	0.9101
1.8253	10.0	830	0.9296	0.5595	0.9430	0.9296
1.7091	11.0	913	0.9451	0.4272	0.9536	0.9451
1.4114	12.0	996	0.9583	0.3359	0.9651	0.9583
1.261	13.0	1079	0.9690	0.2632	0.9736	0.9690
1.107	14.0	1162	0.9718	0.2092	0.9749	0.9718
1.044	15.0	1245	0.9770	0.1753	0.9818	0.9770
0.8997	16.0	1328	0.9858	0.1412	0.9885	0.9858
0.7766	17.0	1411	0.9820	0.1218	0.9857	0.9820
0.7512	18.0	1494	0.9883	0.1019	0.9905	0.9883
0.6487	19.0	1577	0.9913	0.0848	0.9930	0.9913
0.5751	20.0	1660	0.9903	0.0736	0.9919	0.9903
0.596	21.0	1743	0.9915	0.0651	0.9928	0.9915
0.5264	22.0	1826	0.9915	0.0563	0.9935	0.9915
0.6321	23.0	1909	0.9940	0.0528	0.9952	0.9940
0.5068	24.0	1992	0.9925	0.0429	0.9940	0.9925
0.4243	25.0	2075	0.9940	0.0401	0.9951	0.9940
0.4105	26.0	2158	0.9928	0.0373	0.9942	0.9928
0.3813	27.0	2241	0.9953	0.0327	0.9961	0.9953
0.4051	28.0	2324	0.9940	0.0292	0.9955	0.9940
0.3907	29.0	2407	0.9920	0.0317	0.9937	0.9920
0.4021	30.0	2490	0.9955	0.0263	0.9963	0.9955
0.3351	31.0	2573	0.9953	0.0236	0.9963	0.9953
0.3122	32.0	2656	0.9958	0.0238	0.9966	0.9958
0.2766	33.0	2739	0.9973	0.0181	0.9978	0.9973
0.2536	34.0	2822	0.9943	0.0220	0.9952	0.9943
0.4082	35.0	2905	0.9968	0.0161	0.9974	0.9968
0.2783	36.0	2988	0.9965	0.0178	0.9971	0.9965
0.2539	37.0	3071	0.9958	0.0150	0.9967	0.9958
0.214	38.0	3154	0.9965	0.0169	0.9969	0.9965
0.2502	39.0	3237	0.9958	0.0168	0.9964	0.9958
0.2831	40.0	3320	0.9975	0.0123	0.9980	0.9975
0.2116	41.0	3403	0.9978	0.0115	0.9980	0.9978
0.2261	42.0	3486	0.9978	0.0130	0.9982	0.9978
0.2676	43.0	3569	0.9978	0.0153	0.9984	0.9978
0.2331	44.0	3652	0.9960	0.0149	0.9966	0.9960
0.3075	45.0	3735	0.9963	0.0118	0.9970	0.9963
0.1773	46.0	3818	0.9958	0.0127	0.9968	0.9958
0.2338	47.0	3901	0.0088	0.9985	0.9988	0.9985
0.2109	48.0	3984	0.0091	0.9980	0.9985	0.9980

Framework versions

• Transformers 4.44.2
• Pytorch 2.4.1+cu121
• Datasets 3.2.0
• Tokenizers 0.19.1