videomae-base-finetuned-kinetics-0314-clip_duration-abnormal02

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: 0.1692
• Accuracy: 0.9393

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: Use adamw_torch with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: linear
• lr_scheduler_warmup_ratio: 0.05
• training_steps: 37200

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
0.2117	0.0100	373	0.4157	0.8475
0.2524	1.0100	746	0.3192	0.8780
0.0027	2.0100	1119	0.4393	0.8607
0.4521	3.0100	1492	0.7233	0.8037
0.0416	4.0100	1865	0.3474	0.8833
0.0025	5.0100	2238	0.3456	0.8912
0.4299	6.0100	2611	0.2992	0.9032
0.0028	7.0100	2984	0.5732	0.8753
0.002	8.0100	3357	0.4585	0.8846
0.0338	9.0100	3730	0.4494	0.8926
0.0005	10.0100	4103	0.7610	0.8382
0.0008	11.0100	4476	0.5538	0.8846
0.1012	12.0100	4849	0.4533	0.8992
0.0013	13.0100	5222	0.4879	0.8899
0.0002	14.0100	5595	0.5813	0.8621
0.0004	15.0100	5968	0.5061	0.8979
1.0545	16.0100	6341	0.6415	0.8674

Framework versions

• Transformers 4.48.0
• Pytorch 2.5.1+cu124
• Datasets 3.2.0
• Tokenizers 0.21.0