videomae-base-finetuned-subset-100epochs

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: 0.7077
• Accuracy: 0.7685

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: 5e-05
• train_batch_size: 8
• eval_batch_size: 8
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_ratio: 0.1
• training_steps: 5550

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
1.6657	0.01	56	1.6248	0.2258
1.6109	1.01	112	1.5601	0.3917
1.5669	2.01	168	1.5563	0.3733
1.45	3.01	224	1.0988	0.5991
1.1208	4.01	280	1.2279	0.5714
1.1588	5.01	336	0.8424	0.7097
1.0834	6.01	392	1.1035	0.5346
1.2194	7.01	448	1.0749	0.4839
0.8462	8.01	504	0.8755	0.6406
1.058	9.01	560	0.9025	0.6498
1.0163	10.01	616	1.2588	0.4839
1.0639	11.01	672	0.8928	0.6359
0.9317	12.01	728	0.8825	0.6221
0.9038	13.01	784	0.8765	0.5622
0.9155	14.01	840	0.8431	0.7005
1.0731	15.01	896	0.8175	0.7005
0.6864	16.01	952	1.0591	0.5853
0.9537	17.01	1008	0.9703	0.6221
0.7499	18.01	1064	0.8371	0.5806
0.7142	19.01	1120	0.9132	0.6636
0.675	20.01	1176	0.7597	0.6728
0.604	21.01	1232	1.2004	0.5714
0.7738	22.01	1288	1.0633	0.5668
0.7651	23.01	1344	0.6865	0.6820
0.6292	24.01	1400	0.7607	0.6912
0.7387	25.01	1456	1.3038	0.5346
0.7038	26.01	1512	1.2832	0.5530
0.7565	27.01	1568	0.8128	0.7005
0.6516	28.01	1624	1.0893	0.5392
0.7074	29.01	1680	1.0894	0.5991
0.4902	30.01	1736	1.0695	0.5622
0.4563	31.01	1792	1.2922	0.5300
0.7543	32.01	1848	0.8960	0.6820
0.7467	33.01	1904	0.7861	0.7465
0.6459	34.01	1960	1.2835	0.5622
0.7296	35.01	2016	1.0303	0.5806
0.5	36.01	2072	0.8924	0.6129
0.5181	37.01	2128	0.8769	0.7235
0.5225	38.01	2184	0.7288	0.7512
0.5617	39.01	2240	0.6330	0.7926
0.677	40.01	2296	0.7733	0.7419
0.6891	41.01	2352	0.7463	0.8157
0.6662	42.01	2408	0.9304	0.7235
0.4602	43.01	2464	1.5115	0.5207
0.581	44.01	2520	1.2296	0.6175
0.5418	45.01	2576	1.0070	0.6221
0.5199	46.01	2632	1.1344	0.6083
0.6876	47.01	2688	0.9800	0.5760
0.5165	48.01	2744	1.3709	0.5069
0.5727	49.01	2800	0.9960	0.6866
0.3698	50.01	2856	1.2246	0.5484
0.5836	51.01	2912	0.9892	0.6866
0.6017	52.01	2968	0.9388	0.6590
0.4851	53.01	3024	1.1415	0.6590
0.3038	54.01	3080	0.9413	0.6959
0.6075	55.01	3136	1.0467	0.6129
0.4474	56.01	3192	0.8436	0.6866
0.3711	57.01	3248	0.8994	0.6774
0.5279	58.01	3304	0.8859	0.7189
0.6032	59.01	3360	1.2931	0.6498
0.3282	60.01	3416	0.9435	0.7143
0.3506	61.01	3472	1.0971	0.6728
0.3169	62.01	3528	0.9101	0.7512
0.438	63.01	3584	1.4072	0.6359
0.5208	64.01	3640	1.2648	0.6544
0.4563	65.01	3696	1.1162	0.6498
0.6693	66.01	3752	1.8558	0.5576
0.5599	67.01	3808	1.6574	0.5392
0.4751	68.01	3864	1.1883	0.6129
0.6489	69.01	3920	1.2733	0.6129
0.4229	70.01	3976	1.0994	0.6682
0.4194	71.01	4032	1.1464	0.6175
0.2121	72.01	4088	1.1798	0.6175
0.4106	73.01	4144	1.3294	0.5806
0.3962	74.01	4200	1.4209	0.6359
0.2963	75.01	4256	1.5016	0.5945
0.5436	76.01	4312	1.5647	0.5484
0.4115	77.01	4368	1.4309	0.6037
0.1635	78.01	4424	1.3660	0.6452
0.2931	79.01	4480	1.3299	0.6498
0.5154	80.01	4536	1.6550	0.5806
0.2993	81.01	4592	1.6520	0.5991
0.4391	82.01	4648	1.3823	0.6406
0.485	83.01	4704	1.4860	0.6037
0.3313	84.01	4760	1.3875	0.6175
0.4194	85.01	4816	1.4334	0.5899
0.4515	86.01	4872	1.6489	0.5991
0.3283	87.01	4928	1.4549	0.6083
0.1914	88.01	4984	1.3415	0.6267
0.2142	89.01	5040	1.6426	0.6267
0.3121	90.01	5096	1.6999	0.6037
0.367	91.01	5152	1.4683	0.6083
0.178	92.01	5208	1.4665	0.6267
0.3972	93.01	5264	1.3464	0.6452
0.224	94.01	5320	1.5009	0.6175
0.1848	95.01	5376	1.5068	0.6129
0.2776	96.01	5432	1.5383	0.6175
0.3506	97.01	5488	1.5356	0.6129
0.401	98.01	5544	1.5504	0.6175
0.3466	99.0	5550	1.5505	0.6175

Framework versions

• Transformers 4.36.2
• Pytorch 1.13.1
• Datasets 2.16.1
• Tokenizers 0.15.0