videomae-base-finetuned-ucf_crime2

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: 1.8463
• Accuracy: 0.5200

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: 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: 2700

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
0.475	0.05	135	0.9935	0.6004
1.44	1.05	270	1.4196	0.4274
1.1084	2.05	405	0.9135	0.6737
0.8732	3.05	540	1.1984	0.5479
1.4184	4.05	675	1.3373	0.4926
1.1355	5.05	810	0.9888	0.6148
0.4522	6.05	945	1.0745	0.5694
0.7754	7.05	1080	1.5848	0.5330
1.1235	8.05	1215	1.3688	0.5753
1.611	9.05	1350	0.6958	0.7694
0.5714	10.05	1485	0.8027	0.7542
0.716	11.05	1620	1.3503	0.6782
0.6642	12.05	1755	1.0798	0.6957
0.8451	13.05	1890	1.2328	0.7479
0.6157	14.05	2025	1.9403	0.5762
0.3358	15.05	2160	1.3435	0.6939
0.5394	16.05	2295	1.2524	0.7056
0.3334	17.05	2430	1.1190	0.7645
0.3513	18.05	2565	1.2137	0.7461
0.2531	19.05	2700	1.2131	0.7362

Framework versions

• Transformers 4.29.2
• Pytorch 1.13.1+cu117
• Datasets 2.12.0
• Tokenizers 0.13.3