videomae-base-finetuned-kinetics-0313-clip_duration-abnormal09_shortsidescale

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: 0.2084
• Accuracy: 0.9553

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: 31000

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
0.0055	0.0100	311	0.3958	0.8691
0.0753	1.0100	622	0.3683	0.8708
0.0109	2.0100	933	0.3968	0.8643
0.0617	3.0100	1244	0.3480	0.8659
0.1567	4.0100	1555	0.4075	0.8449
0.6073	5.0100	1866	0.3312	0.8788
0.0055	6.0100	2177	0.4617	0.8659
0.0007	7.0100	2488	0.4995	0.8708
0.0145	8.0100	2799	0.5133	0.8174
0.0381	9.0100	3110	0.4475	0.8659
0.2152	10.0100	3421	0.6625	0.8498
0.0032	11.0100	3732	0.5294	0.8643
0.049	12.0100	4043	0.4688	0.8724
0.0004	13.0100	4354	0.5969	0.8611
0.0754	14.0100	4665	0.5157	0.8901
0.0043	15.0100	4976	0.5309	0.8756
0.0001	16.0100	5287	0.6341	0.8756
0.0002	17.0100	5598	0.5637	0.8982
0.0002	18.0100	5909	0.6114	0.8901
0.0002	19.0100	6220	0.5572	0.9015
0.0007	20.0100	6531	0.6986	0.8708
0.0003	21.0100	6842	0.6029	0.8918
0.0056	22.0100	7153	0.6299	0.8901
0.0001	23.0100	7464	0.6450	0.8885
0.0	24.0100	7775	0.6583	0.8918
0.0001	25.0100	8086	0.6455	0.8918
0.0082	26.0100	8397	0.6384	0.8918
0.0001	27.0100	8708	0.6429	0.8853
0.0001	28.0100	9019	0.6509	0.8885
0.0001	29.0100	9330	0.6478	0.8885

Framework versions

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