videomae-base-finetuned-subset-check10

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: 0.8682
• Accuracy: 0.6343

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: 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: 2220

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
1.5175	0.03	56	1.6041	0.2074
1.4397	1.03	112	1.4559	0.3871
1.464	2.03	168	1.3637	0.3963
1.3404	3.03	224	1.2467	0.4470
1.3284	4.03	280	1.3115	0.3318
1.1598	5.03	336	1.2489	0.4470
0.9615	6.03	392	1.3057	0.4009
0.9357	7.03	448	0.9201	0.6498
0.9785	8.03	504	0.8629	0.6774
1.0862	9.03	560	1.0977	0.5069
0.9315	10.03	616	0.7868	0.7097
0.9404	11.03	672	0.8170	0.6728
0.939	12.03	728	0.9246	0.6636
0.8205	13.03	784	0.8420	0.6866
0.6719	14.03	840	1.0725	0.5899
0.8308	15.03	896	0.8683	0.6912
0.7554	16.03	952	0.9684	0.5991
0.6962	17.03	1008	1.1106	0.5484
0.7995	18.03	1064	0.9751	0.6498
0.8298	19.03	1120	1.0631	0.5300
0.6607	20.03	1176	0.9458	0.6175
0.688	21.03	1232	1.0296	0.6037
0.5835	22.03	1288	0.8948	0.6774
0.6987	23.03	1344	0.7883	0.7189
0.4979	24.03	1400	0.7089	0.7189
0.6163	25.03	1456	0.7634	0.7235
0.6754	26.03	1512	0.9444	0.6359
0.6673	27.03	1568	0.8391	0.6544
0.4924	28.03	1624	0.8289	0.6682
0.6438	29.03	1680	0.9605	0.6129
0.5714	30.03	1736	0.8838	0.6452
0.6726	31.03	1792	0.8412	0.6590
0.5027	32.03	1848	0.8439	0.6728
0.4649	33.03	1904	0.9525	0.6267
0.6625	34.03	1960	0.7850	0.7281
0.5793	35.03	2016	0.8481	0.6728
0.6411	36.03	2072	0.8842	0.6590
0.6592	37.03	2128	0.8028	0.6912
0.5524	38.03	2184	0.8216	0.6866
0.5697	39.02	2220	0.8339	0.6774

Framework versions

• Transformers 4.36.2
• Pytorch 2.1.1
• Datasets 2.16.1
• Tokenizers 0.15.0