videomae-base-finetuned-signlanguage

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

• Loss: 1.3358
• Accuracy: 0.6567

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: 1e-05
• train_batch_size: 32
• eval_batch_size: 32
• 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: linear
• lr_scheduler_warmup_ratio: 0.1
• training_steps: 3560

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
4.2372	0.0065	23	4.2439	0.0186
4.2325	1.0065	46	4.2369	0.0186
4.1795	2.0065	69	4.2250	0.0233
4.211	3.0065	92	4.2141	0.0140
4.2361	4.0065	115	4.2051	0.0372
4.2524	5.0065	138	4.2052	0.0233
4.2297	6.0065	161	4.2014	0.0186
4.2596	7.0065	184	4.1950	0.0233
4.1911	8.0065	207	4.1898	0.0326
4.184	9.0065	230	4.1860	0.0279
4.1731	10.0065	253	4.1680	0.0419
4.098	11.0065	276	4.1413	0.0512
4.1105	12.0065	299	4.1351	0.0279
4.1477	13.0065	322	4.1182	0.0372
4.0121	14.0065	345	4.0324	0.0605
3.8587	15.0065	368	3.9973	0.0512
3.8878	16.0065	391	3.9077	0.0651
3.7205	17.0065	414	3.8901	0.0698
3.6613	18.0065	437	3.7347	0.1349
3.5438	19.0065	460	3.6275	0.1488
3.4033	20.0065	483	3.5495	0.1442
3.2043	21.0065	506	3.5864	0.1349
3.1477	22.0065	529	3.4515	0.1674
3.0344	23.0065	552	3.3110	0.2233
2.9459	24.0065	575	3.2645	0.2605
2.6629	25.0065	598	3.1746	0.2558
2.764	26.0065	621	3.0833	0.3163
2.4924	27.0065	644	2.9918	0.3023
2.6696	28.0065	667	3.0009	0.3349
2.4616	29.0065	690	2.8396	0.4
2.2084	30.0065	713	2.8039	0.3674
2.3011	31.0065	736	2.7465	0.4
2.1059	32.0065	759	2.6865	0.4140
2.0525	33.0065	782	2.6070	0.4326
2.1054	34.0065	805	2.6387	0.3953
1.8791	35.0065	828	2.5539	0.4326
1.7834	36.0065	851	2.4750	0.4326
1.5749	37.0065	874	2.4880	0.4233
1.6162	38.0065	897	2.3581	0.4884
1.5611	39.0065	920	2.2846	0.5256
1.5449	40.0065	943	2.2999	0.5116
1.6013	41.0065	966	2.2349	0.5302
1.3959	42.0065	989	2.1957	0.5488
1.1607	43.0065	1012	2.1559	0.5209
1.2663	44.0065	1035	2.1192	0.5721
1.0869	45.0065	1058	2.0522	0.5535
1.2477	46.0065	1081	2.0597	0.5767
0.9665	47.0065	1104	2.0365	0.5535
1.0521	48.0065	1127	1.9669	0.5767
0.8273	49.0065	1150	1.9918	0.5907
0.8396	50.0065	1173	1.9576	0.5860
0.9543	51.0065	1196	1.9371	0.5953
0.8095	52.0065	1219	1.8800	0.5953
0.7694	53.0065	1242	1.8737	0.6
0.8243	54.0065	1265	1.8846	0.6093
0.6632	55.0065	1288	1.8230	0.6
0.7446	56.0065	1311	1.7898	0.6093
0.7044	57.0065	1334	1.7740	0.5907
0.6732	58.0065	1357	1.8061	0.6047
0.5786	59.0065	1380	1.7060	0.6186
0.6348	60.0065	1403	1.7004	0.6140
0.5706	61.0065	1426	1.7013	0.6279
0.5007	62.0065	1449	1.6992	0.6186
0.5078	63.0065	1472	1.6649	0.6047
0.5048	64.0065	1495	1.6449	0.6140
0.4526	65.0065	1518	1.6256	0.6279
0.504	66.0065	1541	1.6401	0.6372
0.3824	67.0065	1564	1.5941	0.6093
0.453	68.0065	1587	1.6236	0.6186
0.3618	69.0065	1610	1.6100	0.6186
0.3689	70.0065	1633	1.5488	0.6419
0.3545	71.0065	1656	1.5390	0.6465
0.4126	72.0065	1679	1.5287	0.6558
0.2734	73.0065	1702	1.4978	0.6465
0.3144	74.0065	1725	1.5038	0.6326
0.3152	75.0065	1748	1.5692	0.6326
0.371	76.0065	1771	1.5331	0.6558
0.3033	77.0065	1794	1.4733	0.6465
0.2574	78.0065	1817	1.5694	0.5907
0.2562	79.0065	1840	1.5097	0.6279
0.2162	80.0065	1863	1.4782	0.6512
0.2493	81.0065	1886	1.4350	0.6465
0.2173	82.0065	1909	1.4730	0.6093
0.2508	83.0065	1932	1.4735	0.6186
0.1932	84.0065	1955	1.4491	0.6326
0.1822	85.0065	1978	1.4155	0.6326
0.2051	86.0065	2001	1.4431	0.6419
0.2269	87.0065	2024	1.4029	0.6419
0.1747	88.0065	2047	1.4643	0.6233
0.1464	89.0065	2070	1.3921	0.6558
0.1642	90.0065	2093	1.4033	0.6512
0.1582	91.0065	2116	1.3728	0.6512
0.1641	92.0065	2139	1.3756	0.6326
0.1292	93.0065	2162	1.3731	0.6512
0.1285	94.0065	2185	1.3559	0.6698
0.1405	95.0065	2208	1.4126	0.6233
0.1299	96.0065	2231	1.3524	0.6419
0.1166	97.0065	2254	1.3812	0.6512
0.1434	98.0065	2277	1.4055	0.6279
0.1748	99.0065	2300	1.3894	0.6558
0.0999	100.0065	2323	1.3665	0.6326
0.1361	101.0065	2346	1.3776	0.6372
0.118	102.0065	2369	1.3635	0.6558
0.0996	103.0065	2392	1.3477	0.6512
0.1232	104.0065	2415	1.3550	0.6419
0.0783	105.0065	2438	1.3460	0.6233
0.1517	106.0065	2461	1.3527	0.6279
0.1007	107.0065	2484	1.3040	0.6465
0.1036	108.0065	2507	1.3216	0.6698
0.1085	109.0065	2530	1.2975	0.6326
0.0691	110.0065	2553	1.3401	0.6512
0.1231	111.0065	2576	1.3251	0.6372
0.0801	112.0065	2599	1.3120	0.6605
0.0784	113.0065	2622	1.3061	0.6605
0.0891	114.0065	2645	1.2882	0.6558
0.0792	115.0065	2668	1.3531	0.6558
0.0772	116.0065	2691	1.3200	0.6698
0.1068	117.0065	2714	1.3186	0.6744
0.0711	118.0065	2737	1.3067	0.6419
0.0982	119.0065	2760	1.3161	0.6512
0.0741	120.0065	2783	1.3029	0.6512
0.1507	121.0065	2806	1.3406	0.6605
0.0602	122.0065	2829	1.3187	0.6558
0.0748	123.0065	2852	1.2874	0.6605
0.0638	124.0065	2875	1.2871	0.6791
0.0915	125.0065	2898	1.2869	0.6465
0.0749	126.0065	2921	1.2859	0.6558
0.0717	127.0065	2944	1.3222	0.6372
0.0539	128.0065	2967	1.3263	0.6326
0.0488	129.0065	2990	1.2945	0.6512
0.0696	130.0065	3013	1.2636	0.6698
0.0665	131.0065	3036	1.2910	0.6698
0.0562	132.0065	3059	1.2820	0.6558
0.0527	133.0065	3082	1.2927	0.6651
0.057	134.0065	3105	1.2846	0.6558
0.0764	135.0065	3128	1.3104	0.6651
0.0805	136.0065	3151	1.3110	0.6465
0.0503	137.0065	3174	1.3107	0.6558
0.0629	138.0065	3197	1.2915	0.6465
0.0491	139.0065	3220	1.2753	0.6558
0.0456	140.0065	3243	1.3105	0.6605
0.053	141.0065	3266	1.2686	0.6558
0.051	142.0065	3289	1.2831	0.6651
0.0669	143.0065	3312	1.2852	0.6698
0.0445	144.0065	3335	1.2868	0.6651
0.0414	145.0065	3358	1.2806	0.6698
0.0575	146.0065	3381	1.2815	0.6651
0.0392	147.0065	3404	1.2596	0.6558
0.0898	148.0065	3427	1.2666	0.6698
0.0617	149.0065	3450	1.2629	0.6605
0.0453	150.0065	3473	1.2797	0.6651
0.0398	151.0065	3496	1.2698	0.6558
0.0435	152.0065	3519	1.2721	0.6558
0.0507	153.0065	3542	1.2745	0.6512
0.0525	154.0051	3560	1.2777	0.6512

Framework versions

• Transformers 4.49.0
• Pytorch 2.0.1+cu118
• Datasets 3.3.2
• Tokenizers 0.21.0