SignBart-KArSL01-502

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

• Loss: 0.0285
• Accuracy: 0.9908
• Precision: 0.9901
• Recall: 0.9908

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
6.6082	1.0	83	0.0211	5.7261	0.0033	0.0211
6.0127	2.0	166	0.1409	4.7101	0.1049	0.1409
5.2488	3.0	249	0.3636	3.8251	0.3684	0.3636
4.6145	4.0	332	0.5485	3.0650	0.5544	0.5485
4.0153	5.0	415	0.6767	2.3960	0.6925	0.6767
3.5522	6.0	498	0.7838	1.8625	0.7987	0.7838
3.0186	7.0	581	0.8415	1.4073	0.8521	0.8415
2.6805	8.0	664	0.8740	1.0787	0.8871	0.8740
2.3229	9.0	747	0.9011	0.8302	0.9127	0.9011
1.9478	10.0	830	0.9299	0.6075	0.9403	0.9299
1.7953	11.0	913	0.9411	0.4838	0.9478	0.9411
1.5113	12.0	996	0.9535	0.3823	0.9589	0.9535
1.3588	13.0	1079	0.9622	0.3039	0.9657	0.9622
1.1982	14.0	1162	0.9622	0.2500	0.9645	0.9622
1.1369	15.0	1245	0.9687	0.2053	0.9714	0.9687
1.0129	16.0	1328	0.9704	0.1764	0.9707	0.9704
0.8517	17.0	1411	0.9694	0.1542	0.9719	0.9694
0.8642	18.0	1494	0.9781	0.1306	0.9800	0.9781
0.7324	19.0	1577	0.9819	0.1158	0.9829	0.9819
0.6718	20.0	1660	0.9759	0.1061	0.9772	0.9759
0.6456	21.0	1743	0.9816	0.0922	0.9824	0.9816
0.5705	22.0	1826	0.9819	0.0844	0.9844	0.9819
0.6644	23.0	1909	0.9779	0.0804	0.9783	0.9779
0.5754	24.0	1992	0.9853	0.0696	0.9857	0.9853
0.4832	25.0	2075	0.9843	0.0685	0.9854	0.9843
0.4655	26.0	2158	0.9824	0.0634	0.9821	0.9824
0.4585	27.0	2241	0.9861	0.0566	0.9859	0.9861
0.4404	28.0	2324	0.9848	0.0567	0.9839	0.9848
0.4272	29.0	2407	0.9863	0.0522	0.9879	0.9863
0.4411	30.0	2490	0.9871	0.0496	0.9893	0.9871
0.3749	31.0	2573	0.9903	0.0443	0.9912	0.9903
0.3564	32.0	2656	0.9866	0.0430	0.9885	0.9866
0.3425	33.0	2739	0.9906	0.0441	0.9894	0.9906
0.3413	34.0	2822	0.9881	0.0399	0.9898	0.9881
0.4125	35.0	2905	0.9871	0.0440	0.9864	0.9871
0.2825	36.0	2988	0.9881	0.0375	0.9886	0.9881
0.3285	37.0	3071	0.9903	0.0369	0.9894	0.9903
0.2496	38.0	3154	0.9901	0.0367	0.9892	0.9901
0.2859	39.0	3237	0.9891	0.0398	0.9884	0.9891
0.2653	40.0	3320	0.9888	0.0421	0.9886	0.9888
0.2565	41.0	3403	0.9886	0.0377	0.9895	0.9886
0.2674	42.0	3486	0.9866	0.0348	0.9881	0.9866
0.3163	43.0	3569	0.9888	0.0322	0.9895	0.9888
0.2357	44.0	3652	0.9906	0.0302	0.9900	0.9906
0.3155	45.0	3735	0.9896	0.0328	0.9899	0.9896
0.2068	46.0	3818	0.9878	0.0357	0.9878	0.9878
0.2536	47.0	3901	0.9888	0.0304	0.9890	0.9888
0.2035	48.0	3984	0.9888	0.0352	0.9882	0.9888
0.2385	49.0	4067	0.9888	0.0355	0.9902	0.9888
0.2702	50.0	4150	0.0285	0.9908	0.9901	0.9908

Framework versions

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