VideoMAE_Kinetics_fold__3__BdSLW60_SKF

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: 0.0284
• Accuracy: 0.9963

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
• gradient_accumulation_steps: 4
• total_train_batch_size: 8
• 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: 0.1
• training_steps: 9030
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
8.4384	0.1	903	0.3165	0.9440
0.3934	1.1001	1807	0.0428	0.9851
0.0771	2.1	2710	0.0484	0.9913
0.0486	3.1001	3614	0.0758	0.9875
0.0303	4.1	4517	0.0677	0.9900
0.0316	5.1001	5421	0.0564	0.9888
0.0263	6.1	6324	0.0408	0.9950
0.0076	7.1001	7228	0.0525	0.9913
0.0135	8.1	8131	0.0340	0.9963
0.007	9.0995	9030	0.0284	0.9963

Framework versions

• Transformers 4.46.1
• Pytorch 2.5.1+cu124
• Datasets 3.1.0
• Tokenizers 0.20.1