SignBart-WLASL-1000

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

• Loss: 1.1030
• Accuracy: 0.7625
• Precision: 0.7031
• Recall: 0.7625

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: Use adamw_torch with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: cosine
• num_epochs: 1000

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	Precision	Recall
6.9826	1.0	55	6.8978	0.001	0.0000	0.001
6.8146	2.0	110	6.6617	0.005	0.0001	0.005
6.5947	3.0	165	6.4332	0.0105	0.0028	0.0105
6.3883	4.0	220	6.1928	0.015	0.0046	0.015
6.183	5.0	275	5.9518	0.032	0.0107	0.032
5.9609	6.0	330	5.7551	0.0425	0.0137	0.0425
5.7581	7.0	385	5.5601	0.058	0.0206	0.058
5.6294	8.0	440	5.3768	0.0915	0.0346	0.0915
5.4254	9.0	495	5.2025	0.107	0.0453	0.107
5.2822	10.0	550	5.0386	0.1275	0.0617	0.1275
5.102	11.0	605	4.9017	0.1515	0.0846	0.1515
4.9978	12.0	660	4.7400	0.1605	0.0841	0.1605
4.8376	13.0	715	4.5917	0.1945	0.1090	0.1945
4.7316	14.0	770	4.4440	0.215	0.1278	0.215
4.5808	15.0	825	4.3048	0.234	0.1420	0.234
4.5496	16.0	880	4.1654	0.265	0.1680	0.265
4.3108	17.0	935	4.0429	0.273	0.1749	0.273
4.2491	18.0	990	3.9161	0.301	0.1991	0.301
4.1709	19.0	1045	3.7963	0.328	0.2152	0.328
4.1081	20.0	1100	3.6928	0.339	0.2318	0.339
3.9038	21.0	1155	3.5786	0.3545	0.2503	0.3545
3.8896	22.0	1210	3.4812	0.374	0.2678	0.374
3.7007	23.0	1265	3.3867	0.3765	0.2729	0.3765
3.6388	24.0	1320	3.2732	0.3945	0.2897	0.3945
3.4366	25.0	1375	3.1881	0.4035	0.2968	0.4035
3.3915	26.0	1430	3.1018	0.4195	0.3109	0.4195
3.3078	27.0	1485	3.0161	0.433	0.3269	0.433
3.3424	28.0	1540	2.9368	0.4505	0.3395	0.4505
3.1481	29.0	1595	2.8653	0.447	0.3407	0.447
3.1006	30.0	1650	2.7771	0.4635	0.3557	0.4635
3.0402	31.0	1705	2.7295	0.476	0.3746	0.476
2.9198	32.0	1760	2.6286	0.4955	0.3911	0.4955
2.9306	33.0	1815	2.5719	0.4955	0.3905	0.4955
2.7992	34.0	1870	2.4924	0.4995	0.3978	0.4995
2.8016	35.0	1925	2.4390	0.512	0.4109	0.512
2.7209	36.0	1980	2.4022	0.5235	0.4176	0.5235
2.6578	37.0	2035	2.3125	0.532	0.4290	0.532
2.617	38.0	2090	2.2881	0.5335	0.4354	0.5335
2.4782	39.0	2145	2.2154	0.537	0.4375	0.537
2.4105	40.0	2200	2.1513	0.5565	0.4509	0.5565
2.3521	41.0	2255	2.1075	0.575	0.4740	0.575
2.4844	42.0	2310	2.0762	0.57	0.4764	0.57
2.3568	43.0	2365	2.0327	0.578	0.4806	0.578
2.3927	44.0	2420	1.9975	0.5795	0.4845	0.5795
2.0977	45.0	2475	1.9511	0.5845	0.4924	0.5845
2.0484	46.0	2530	1.9507	0.5785	0.4866	0.5785
2.1581	47.0	2585	1.8796	0.5965	0.5072	0.5965
2.0352	48.0	2640	1.8491	0.5965	0.5045	0.5965
2.0802	49.0	2695	1.8200	0.616	0.5281	0.616
1.8752	50.0	2750	1.8062	0.6085	0.5213	0.6085
1.9879	51.0	2805	1.7768	0.622	0.5372	0.622
2.0026	52.0	2860	1.7399	0.616	0.5331	0.616
1.9185	53.0	2915	1.7104	0.632	0.5525	0.632
1.8116	54.0	2970	1.6671	0.6305	0.5450	0.6305
1.8589	55.0	3025	1.6380	0.63	0.5465	0.63
1.7824	56.0	3080	1.6134	0.6545	0.5781	0.6545
1.6361	57.0	3135	1.6197	0.6355	0.5522	0.6355
1.5969	58.0	3190	1.5865	0.647	0.5709	0.647
1.7796	59.0	3245	1.5638	0.6535	0.5830	0.6535
1.6418	60.0	3300	1.5520	0.654	0.5821	0.654
1.6398	61.0	3355	1.5422	0.646	0.5662	0.646
1.7401	62.0	3410	1.5019	0.667	0.5915	0.667
1.5695	63.0	3465	1.5018	0.6555	0.5805	0.6555
1.7614	64.0	3520	1.4665	0.661	0.5831	0.661
1.7522	65.0	3575	1.4738	0.667	0.5882	0.667
1.4649	66.0	3630	1.4502	0.67	0.5922	0.67
1.4873	67.0	3685	1.4433	0.666	0.5915	0.666
1.5402	68.0	3740	1.4289	0.6775	0.6009	0.6775
1.4602	69.0	3795	1.3836	0.6885	0.6148	0.6885
1.4702	70.0	3850	1.3791	0.6865	0.6124	0.6865
1.4524	71.0	3905	1.3544	0.689	0.6171	0.689
1.352	72.0	3960	1.3390	0.7015	0.6272	0.7015
1.398	73.0	4015	1.3304	0.686	0.6155	0.686
1.4414	74.0	4070	1.3008	0.706	0.6315	0.706
1.2863	75.0	4125	1.3093	0.7025	0.6346	0.7025
1.4229	76.0	4180	1.2865	0.7095	0.6329	0.7095
1.4661	77.0	4235	1.3228	0.6935	0.6274	0.6935
1.2171	78.0	4290	1.2663	0.7205	0.6506	0.7205
1.2461	79.0	4345	1.2699	0.718	0.6493	0.718
1.2544	80.0	4400	1.2703	0.722	0.6487	0.722
1.4058	81.0	4455	1.2496	0.729	0.6591	0.729
1.3684	82.0	4510	1.2580	0.712	0.6437	0.712
1.3059	83.0	4565	1.2301	0.719	0.6479	0.719
1.1517	84.0	4620	1.1985	0.72	0.6553	0.72
1.2521	85.0	4675	1.2462	0.7175	0.6471	0.7175
1.1977	86.0	4730	1.2272	0.726	0.6567	0.726
1.2315	87.0	4785	1.2026	0.7295	0.6621	0.7295
1.0442	88.0	4840	1.1870	0.741	0.6755	0.741
1.281	89.0	4895	1.1861	0.7355	0.6627	0.7355
1.1566	90.0	4950	1.1943	0.734	0.6661	0.734
1.1915	91.0	5005	1.1853	0.7355	0.6681	0.7355
1.1377	92.0	5060	1.1974	0.7335	0.6677	0.7335
1.0245	93.0	5115	1.1540	0.7375	0.6738	0.7375
1.187	94.0	5170	1.1842	0.737	0.6768	0.737
1.2676	95.0	5225	1.1525	0.7445	0.6791	0.7445
1.2196	96.0	5280	1.1667	0.7445	0.6816	0.7445
1.0677	97.0	5335	1.1668	0.7465	0.6821	0.7465
1.1943	98.0	5390	1.1738	0.7435	0.6778	0.7435
1.237	99.0	5445	1.1436	0.7485	0.6910	0.7485
1.1528	100.0	5500	1.1405	0.753	0.6937	0.753
1.0181	101.0	5555	1.1456	0.748	0.6862	0.748
1.0136	102.0	5610	1.1312	0.754	0.6905	0.754
1.1207	103.0	5665	1.1086	0.762	0.6975	0.762
1.1906	104.0	5720	1.1508	0.7455	0.6811	0.7455
1.1937	105.0	5775	1.1227	0.7555	0.6939	0.7555
1.209	106.0	5830	1.1152	0.752	0.6907	0.752
1.0693	107.0	5885	1.0996	0.76	0.7011	0.76
1.059	108.0	5940	1.0907	0.769	0.7070	0.769
1.1382	109.0	5995	1.1168	0.758	0.6954	0.758
1.0439	110.0	6050	1.1145	0.762	0.6981	0.762
1.2116	111.0	6105	1.1086	0.7645	0.7040	0.7645
1.0528	112.0	6160	1.1151	0.759	0.6954	0.759
1.1406	113.0	6215	1.1030	0.7625	0.7031	0.7625

Framework versions

• Transformers 4.47.0
• Pytorch 2.5.1+cu121
• Datasets 3.2.0
• Tokenizers 0.21.0