videomae-base-finetuned-subset

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: 1.9422
• Accuracy: 0.7011

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: 2
• eval_batch_size: 2
• 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: 2640

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
1.9136	0.0254	67	2.1529	0.1096
1.9987	1.0254	134	1.9985	0.2593
1.0754	2.0254	201	2.1637	0.1956
1.413	3.0254	268	1.9523	0.1473
1.3047	4.0254	335	2.5442	0.1956
1.2488	5.0254	402	2.0563	0.2330
1.0581	6.0254	469	1.9954	0.2264
0.9165	7.0254	536	1.7661	0.2769
1.5722	8.0254	603	2.2872	0.2264
0.9083	9.0254	670	2.1004	0.3363
0.8093	10.0254	737	1.2497	0.6549
0.4925	11.0254	804	3.5720	0.2813
0.4573	12.0254	871	1.5213	0.3604
1.1082	13.0254	938	1.5453	0.5934
0.8066	14.0254	1005	2.9169	0.2967
0.6615	15.0254	1072	2.1412	0.5780
0.146	16.0254	1139	2.5006	0.3978
0.3815	17.0254	1206	1.7907	0.5956
0.2124	18.0254	1273	1.6622	0.6527
0.5304	19.0254	1340	1.8988	0.5956
0.1519	20.0254	1407	2.7940	0.3934
0.486	21.0254	1474	2.6766	0.4198
0.5502	22.0254	1541	2.3451	0.5495
0.7527	23.0254	1608	1.7518	0.6462
0.3194	24.0254	1675	2.0738	0.5890
0.0189	25.0254	1742	2.9264	0.5407
0.2928	26.0254	1809	2.5495	0.5451
0.0036	27.0254	1876	1.8143	0.6989
0.3772	28.0254	1943	2.2384	0.6088
0.0044	29.0254	2010	1.7688	0.7033
0.7291	30.0254	2077	2.0591	0.6571
0.1553	31.0254	2144	2.0690	0.6505
0.5454	32.0254	2211	1.8762	0.7055
0.3096	33.0254	2278	2.2310	0.6440
0.0053	34.0254	2345	2.0907	0.6615
0.0024	35.0254	2412	2.4127	0.6022
0.0022	36.0254	2479	2.0037	0.6989
0.0026	37.0254	2546	2.0130	0.6725
0.0013	38.0254	2613	1.9391	0.6967
0.0017	39.0102	2640	1.9422	0.7011

Framework versions

• Transformers 4.40.2
• Pytorch 2.0.1+cu117
• Datasets 2.20.0
• Tokenizers 0.19.1