vivit-b-16x2-kinetics400-finetuned-cctv-surveillance

This model is a fine-tuned version of google/vivit-b-16x2-kinetics400 on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 0.1690
• Accuracy: 0.9559
• F1: 0.9430
• Recall: 0.9559
• Precision: 0.9333

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-06
• 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: 4032

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	F1	Recall	Precision
1.5836	0.12	504	0.3644	0.9206	0.8850	0.9206	0.8799
0.3767	1.12	1008	0.2586	0.9265	0.8994	0.9265	0.8831
0.2063	2.12	1512	0.2190	0.9294	0.9097	0.9294	0.9002
0.4514	3.12	2016	0.2217	0.9529	0.9419	0.9529	0.9380
0.2678	4.12	2520	0.1919	0.9529	0.9419	0.9529	0.9380
0.2311	5.12	3024	0.1797	0.9412	0.9252	0.9412	0.9141
0.5256	6.12	3528	0.1690	0.9559	0.9430	0.9559	0.9333
0.539	7.12	4032	0.1678	0.9529	0.9398	0.9529	0.9297

Framework versions

• Transformers 4.39.3
• Pytorch 2.1.2
• Datasets 2.18.0
• Tokenizers 0.15.2