videomae-finetuned

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

• Loss: 0.1289
• Accuracy: 0.9706

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: 16
• eval_batch_size: 16
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_ratio: 0.2
• training_steps: 8620

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
0.7657	0.05	432	0.6104	0.8243
0.4011	1.05	864	0.4283	0.8896
0.301	2.05	1296	0.3957	0.8833
0.2967	3.05	1728	0.3066	0.9138
0.229	4.05	2160	0.3268	0.9092
0.2185	5.05	2592	0.2865	0.9281
0.1863	6.05	3024	0.2550	0.9320
0.1386	7.05	3456	0.2366	0.9422
0.1542	8.05	3888	0.2162	0.9449
0.1396	9.05	4320	0.2069	0.9424
0.1323	10.05	4752	0.2021	0.9486
0.1008	11.05	5184	0.2408	0.9465
0.0756	12.05	5616	0.2023	0.9509
0.0628	13.05	6048	0.1730	0.9609
0.0747	14.05	6480	0.1747	0.9622
0.047	15.05	6912	0.1698	0.9629
0.0586	16.05	7344	0.1412	0.9668
0.0644	17.05	7776	0.1314	0.9704
0.0521	18.05	8208	0.1343	0.9688
0.0516	19.05	8620	0.1289	0.9706

Framework versions

• Transformers 4.38.1
• Pytorch 2.0.1+cu118
• Datasets 2.18.0
• Tokenizers 0.15.2