videomae-base-finetuned-ivt-subset-30

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: 0.3996
• Accuracy: 0.8163

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: 2970

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
0.4176	0.0333	99	0.3317	0.8980
0.394	1.0333	198	0.4196	0.7857
0.4876	2.0333	297	0.2667	0.9184
0.3688	3.0333	396	0.3714	0.8367
0.3578	4.0333	495	0.4811	0.7143
0.1275	5.0333	594	0.8828	0.7245
0.3909	6.0333	693	0.3703	0.8776
0.1795	7.0333	792	0.2431	0.8878
0.1892	8.0333	891	0.3675	0.8878
0.2578	9.0333	990	0.7125	0.8061
0.1941	10.0333	1089	0.3057	0.9184
0.0745	11.0333	1188	0.6525	0.8367
0.041	12.0333	1287	0.4668	0.9082
0.0918	13.0333	1386	0.4170	0.9082
0.1347	14.0333	1485	1.1137	0.7857
0.0712	15.0333	1584	0.5400	0.8878
0.059	16.0333	1683	0.5773	0.8776
0.0011	17.0333	1782	0.4739	0.9082
0.0277	18.0333	1881	0.4686	0.9082
0.039	19.0333	1980	0.5497	0.9082
0.0005	20.0333	2079	0.5712	0.8980
0.0004	21.0333	2178	0.5519	0.8980
0.0003	22.0333	2277	0.5531	0.9082
0.0003	23.0333	2376	0.5200	0.9082
0.0006	24.0333	2475	0.5981	0.9082
0.0031	25.0333	2574	0.5792	0.8980
0.0364	26.0333	2673	0.6266	0.8878
0.0135	27.0333	2772	0.5981	0.9082
0.0003	28.0333	2871	0.5625	0.9082
0.0002	29.0333	2970	0.5668	0.9082

Framework versions

• Transformers 4.44.0
• Pytorch 2.4.0+cu121
• Datasets 2.20.0
• Tokenizers 0.19.1