videomae-base-finetuned-kinetics-0415_final_5sec_org_ab7_val_inside_train_04_except259

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: 0.4084
• Accuracy: 0.8860

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: 65100

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
0.004	0.0100	652	0.3930	0.8818
0.471	1.0100	1304	0.3517	0.8982
0.003	2.0100	1956	0.6755	0.8207
0.2142	3.0100	2608	0.4885	0.8974
0.2487	4.0100	3260	0.6298	0.8575
0.0011	5.0100	3912	0.5771	0.8520
0.0004	6.0100	4564	0.6322	0.8583
0.0048	7.0100	5216	0.3840	0.8778
0.5816	8.0100	5868	0.7402	0.8426
0.0274	9.0100	6520	0.6222	0.8551
0.0002	10.0100	7172	0.6347	0.8872
0.0006	11.0100	7824	0.5242	0.8857
0.0001	12.0100	8476	1.1792	0.8058
0.0006	13.0100	9128	0.8387	0.8449
0.0003	14.0100	9780	0.4948	0.8872
0.0003	15.0100	10432	0.5446	0.8865
0.0008	16.0100	11084	0.7601	0.8583

Framework versions

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