SignBart-KArSL01-190

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

• Loss: 0.0745
• Accuracy: 0.9804
• Precision: 0.9803
• Recall: 0.9804

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
• lr_scheduler_warmup_ratio: 0.1
• num_epochs: 200

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	Precision	Recall
5.1095	1.0	32	5.0121	0.0563	0.0286	0.0563
5.0804	2.0	64	4.9582	0.0694	0.0422	0.0694
5.0279	3.0	96	4.8680	0.0759	0.0543	0.0759
4.9388	4.0	128	4.7387	0.1198	0.0702	0.1198
4.8028	5.0	160	4.5633	0.1721	0.1245	0.1721
4.6544	6.0	192	4.3603	0.2421	0.2145	0.2421
4.4744	7.0	224	4.1342	0.2958	0.2596	0.2958
4.3127	8.0	256	3.9029	0.3593	0.3626	0.3593
4.0826	9.0	288	3.6307	0.4424	0.4670	0.4424
3.8364	10.0	320	3.3442	0.5	0.5212	0.5
3.6191	11.0	352	3.0548	0.5589	0.5668	0.5589
3.3142	12.0	384	2.7417	0.6348	0.6367	0.6348
3.0637	13.0	416	2.4299	0.7016	0.7165	0.7016
2.8353	14.0	448	2.1339	0.7382	0.7548	0.7382
2.6483	15.0	480	1.8488	0.7880	0.8039	0.7880
2.2785	16.0	512	1.5609	0.8331	0.8436	0.8331
2.1246	17.0	544	1.3511	0.8573	0.8708	0.8573
1.9003	18.0	576	1.1286	0.8966	0.9014	0.8966
1.7155	19.0	608	0.9489	0.9208	0.9186	0.9208
1.4837	20.0	640	0.8048	0.9247	0.9241	0.9247
1.5267	21.0	672	0.7032	0.9287	0.9213	0.9287
1.1747	22.0	704	0.5959	0.9437	0.9425	0.9437
1.3094	23.0	736	0.5279	0.9424	0.9382	0.9424
1.109	24.0	768	0.4569	0.9614	0.9553	0.9614
1.0323	25.0	800	0.4127	0.9542	0.9569	0.9542
0.8904	26.0	832	0.3565	0.9627	0.9631	0.9627
1.0299	27.0	864	0.3207	0.9679	0.9689	0.9679
0.8683	28.0	896	0.3005	0.9620	0.9607	0.9620
0.8611	29.0	928	0.2692	0.9653	0.9679	0.9653
0.8391	30.0	960	0.2430	0.9712	0.9680	0.9712
0.8351	31.0	992	0.2442	0.9640	0.9609	0.9640
0.6708	32.0	1024	0.2318	0.9692	0.9680	0.9692
0.7291	33.0	1056	0.2040	0.9732	0.9649	0.9732
0.7557	34.0	1088	0.1995	0.9692	0.9686	0.9692
0.6704	35.0	1120	0.1918	0.9712	0.9719	0.9712
0.6591	36.0	1152	0.1754	0.9725	0.9737	0.9725
0.5812	37.0	1184	0.1722	0.9666	0.9622	0.9666
0.6077	38.0	1216	0.1635	0.9673	0.9708	0.9673
0.5672	39.0	1248	0.1470	0.9758	0.9723	0.9758
0.59	40.0	1280	0.1386	0.9732	0.9765	0.9732
0.566	41.0	1312	0.1285	0.9738	0.9723	0.9738
0.5015	42.0	1344	0.1264	0.9751	0.9746	0.9751
0.6066	43.0	1376	0.1182	0.9771	0.9795	0.9771
0.6864	44.0	1408	0.1322	0.9719	0.9778	0.9719
0.5433	45.0	1440	0.1218	0.9791	0.9797	0.9791
0.5117	46.0	1472	0.1168	0.9745	0.9775	0.9745
0.5838	47.0	1504	0.1155	0.9732	0.9722	0.9732
0.4944	48.0	1536	0.1073	0.9784	0.9828	0.9784
0.4707	49.0	1568	0.0999	0.9791	0.9800	0.9791
0.3833	50.0	1600	0.1055	0.9758	0.9788	0.9758
0.4032	51.0	1632	0.0998	0.9784	0.9826	0.9784
0.5129	52.0	1664	0.1016	0.9764	0.9779	0.9764
0.4409	53.0	1696	0.0960	0.9777	0.9747	0.9777
0.3984	54.0	1728	0.0903	0.9797	0.9831	0.9797
0.3082	55.0	1760	0.0844	0.9797	0.9848	0.9797
0.3639	56.0	1792	0.0817	0.9791	0.9831	0.9791
0.3432	57.0	1824	0.0844	0.9784	0.9775	0.9784
0.3426	58.0	1856	0.0879	0.9777	0.9822	0.9777
0.4153	59.0	1888	0.0782	0.9810	0.9811	0.9810
0.3299	60.0	1920	0.0663	0.9817	0.9845	0.9817
0.4685	61.0	1952	0.0733	0.9823	0.9849	0.9823
0.3758	62.0	1984	0.0754	0.9823	0.9851	0.9823
0.3348	63.0	2016	0.0703	0.9823	0.9836	0.9823
0.4494	64.0	2048	0.0669	0.9823	0.9853	0.9823
0.3787	65.0	2080	0.0745	0.9804	0.9803	0.9804

Framework versions

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