VideoMAE_BdSLW60_FrameRateCorrected_withoutAug_20_epoch_batch_2_2_4

This model is a fine-tuned version of MCG-NJU/videomae-base on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 1.0888
• Accuracy: 0.7704

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: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_ratio: 0.1
• training_steps: 17940
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
2.5781	0.05	897	1.7219	0.6953
0.2879	1.0500	1795	0.1873	0.9624
0.0246	2.0500	2693	0.0435	0.9906
0.0293	3.0500	3591	0.0520	0.9871
0.0811	4.05	4488	0.1336	0.9624
0.0573	5.0500	5386	0.0864	0.9753
0.0261	6.0500	6284	0.1303	0.9765
0.0101	7.0500	7182	0.0963	0.9835

Framework versions

• Transformers 4.44.2
• Pytorch 2.4.1+cu121
• Datasets 2.19.2
• Tokenizers 0.19.1