videomae-base-finetuned-kinetics-0403_final_5sec

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: 0.2477
• Accuracy: 0.9053

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: 64400

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
0.1532	0.0100	645	0.2212	0.9151
0.0051	1.0100	1290	0.8680	0.7644
0.2625	2.0100	1935	0.4553	0.8616
0.0155	3.0100	2580	0.4682	0.8643
0.021	4.0100	3225	0.6284	0.8196
0.6406	5.0100	3870	0.3939	0.9116
1.0417	6.0100	4515	1.1531	0.7898
0.0027	7.0100	5160	0.7326	0.8152
0.0017	8.0100	5805	0.5727	0.8809
0.0009	9.0100	6450	0.4952	0.8949
0.0077	10.0100	7095	0.7302	0.8371
0.0011	11.0100	7740	0.8304	0.8100
0.0002	12.0100	8385	0.8539	0.8135
0.0002	13.0100	9030	1.4690	0.7434
0.9484	14.0100	9675	0.7860	0.8319
0.0145	15.0100	10320	1.3528	0.7758

Framework versions

• Transformers 4.48.1
• Pytorch 2.5.1+cu121
• Datasets 3.2.0
• Tokenizers 0.21.0