SignBart-KArSL-ALL-502

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

• Loss: 0.0253
• Accuracy: 0.9936
• Precision: 0.9941
• Recall: 0.9936

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	Validation Loss	Accuracy	Precision	Recall
6.075	1.0	248	4.2222	0.2192	0.2353	0.2192
4.452	2.0	496	2.5647	0.5864	0.6569	0.5864
3.2856	3.0	744	1.5302	0.7551	0.7884	0.7551
2.5179	4.0	992	0.9342	0.8522	0.8739	0.8522
1.8877	5.0	1240	0.6079	0.8908	0.9064	0.8908
1.5441	6.0	1488	0.4360	0.9177	0.9301	0.9177
1.2167	7.0	1736	0.3084	0.9402	0.9468	0.9402
1.0845	8.0	1984	0.2421	0.9511	0.9576	0.9511
0.9543	9.0	2232	0.2023	0.9551	0.9612	0.9551
0.8612	10.0	2480	0.1635	0.9641	0.9688	0.9641
0.7338	11.0	2728	0.1362	0.9699	0.9732	0.9699
0.667	12.0	2976	0.1167	0.9729	0.9756	0.9729
0.5813	13.0	3224	0.1046	0.9741	0.9778	0.9741
0.5801	14.0	3472	0.0921	0.9782	0.9806	0.9782
0.587	15.0	3720	0.0835	0.9816	0.9839	0.9816
0.5219	16.0	3968	0.0813	0.9795	0.9818	0.9795
0.5142	17.0	4216	0.0737	0.9805	0.9827	0.9805
0.4577	18.0	4464	0.0640	0.9816	0.9836	0.9816
0.4484	19.0	4712	0.0592	0.9856	0.9865	0.9856
0.4431	20.0	4960	0.0599	0.9826	0.9853	0.9826
0.4092	21.0	5208	0.0536	0.9856	0.9869	0.9856
0.432	22.0	5456	0.0469	0.9885	0.9897	0.9885
0.392	23.0	5704	0.0504	0.9863	0.9876	0.9863
0.4146	24.0	5952	0.0484	0.9871	0.9882	0.9871
0.3789	25.0	6200	0.0464	0.9876	0.9889	0.9876
0.3415	26.0	6448	0.0486	0.9867	0.9878	0.9867
0.3605	27.0	6696	0.0437	0.9874	0.9884	0.9874
0.3793	28.0	6944	0.0373	0.9907	0.9915	0.9907
0.3307	29.0	7192	0.0408	0.9884	0.9893	0.9884
0.3288	30.0	7440	0.0347	0.9905	0.9910	0.9905
0.3532	31.0	7688	0.0355	0.9905	0.9915	0.9905
0.3305	32.0	7936	0.0357	0.9899	0.9908	0.9899
0.3275	33.0	8184	0.0360	0.9894	0.9906	0.9894
0.2876	34.0	8432	0.0344	0.9900	0.9908	0.9900
0.2807	35.0	8680	0.0331	0.9912	0.9920	0.9912
0.3086	36.0	8928	0.0325	0.9912	0.9920	0.9912
0.3106	37.0	9176	0.0338	0.9911	0.9918	0.9911
0.3355	38.0	9424	0.0307	0.9911	0.9919	0.9911
0.2987	39.0	9672	0.0348	0.9899	0.9910	0.9899
0.2808	40.0	9920	0.0351	0.9889	0.9898	0.9889
0.2674	41.0	10168	0.0314	0.9910	0.9918	0.9910
0.264	42.0	10416	0.0304	0.9914	0.9922	0.9914
0.282	43.0	10664	0.0299	0.9924	0.9930	0.9924
0.2632	44.0	10912	0.0301	0.9912	0.9920	0.9912
0.2877	45.0	11160	0.0317	0.9906	0.9917	0.9906
0.2692	46.0	11408	0.0253	0.9934	0.9939	0.9934
0.2689	47.0	11656	0.0286	0.9929	0.9935	0.9929
0.2846	48.0	11904	0.0283	0.9917	0.9925	0.9917
0.2614	49.0	12152	0.0323	0.9910	0.9918	0.9910
0.3111	50.0	12400	0.0249	0.9924	0.9931	0.9924
0.2155	51.0	12648	0.0250	0.9934	0.9939	0.9934
0.242	52.0	12896	0.0314	0.9902	0.9910	0.9902
0.2278	53.0	13144	0.0248	0.9929	0.9936	0.9929
0.249	54.0	13392	0.0265	0.9925	0.9932	0.9925
0.2464	55.0	13640	0.0266	0.9929	0.9936	0.9929
0.2225	56.0	13888	0.0224	0.9939	0.9943	0.9939
0.26	57.0	14136	0.0231	0.9944	0.9947	0.9944
0.2298	58.0	14384	0.0262	0.9927	0.9932	0.9927
0.2365	59.0	14632	0.0258	0.9932	0.9939	0.9932
0.1981	60.0	14880	0.0236	0.9931	0.9937	0.9931
0.2062	61.0	15128	0.0253	0.9936	0.9941	0.9936

Framework versions

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