VideoMAE_BdSLW60_FrameRateCorrected_withoutAug

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: 0.7568
• Accuracy: 0.8080

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: 22400

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
3.1009	0.0400	897	2.3993	0.5082
0.3152	1.0401	1795	0.2117	0.9612
0.0363	2.0401	2693	0.0375	0.9918
0.0431	3.0401	3591	0.0841	0.9765
0.0554	4.0400	4488	0.0323	0.9941
0.0467	5.0401	5386	0.0614	0.9835
0.0324	6.0401	6284	0.0944	0.9859
0.0381	7.0401	7182	0.0583	0.9871
0.0039	8.0400	8079	0.0307	0.9918
0.0149	9.0401	8977	0.0694	0.9859

Framework versions

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