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

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

• Loss: 0.1478
• Accuracy: 0.9460
• F1: 0.9430
• Recall: 0.9460
• Precision: 0.9454

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: 4176

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	F1	Recall	Precision
0.9564	0.12	522	0.4417	0.8685	0.8096	0.8685	0.7990
0.4574	1.12	1044	0.2633	0.9131	0.9042	0.9131	0.9269
0.421	2.12	1566	0.1875	0.9272	0.9100	0.9272	0.9353
0.4785	3.12	2088	0.1854	0.9249	0.9082	0.9249	0.9140
0.3213	4.12	2610	0.1805	0.9272	0.9125	0.9272	0.9216
0.1465	5.12	3132	0.1733	0.9413	0.9362	0.9413	0.9398
0.0784	6.12	3654	0.1616	0.9437	0.9391	0.9437	0.9434
0.2017	7.12	4176	0.1478	0.9460	0.9430	0.9460	0.9454

Framework versions

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