VideoMAE_LSA64_SR_12

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: 0.0353
• Accuracy: 0.9922
• Precision: 0.9938
• Recall: 0.9922
• F1: 0.9921

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: 8640
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	Precision	Recall	F1
16.1868	0.0333	288	3.7228	0.0391	0.0138	0.0391	0.0127
13.4364	1.0333	576	2.1449	0.4688	0.4296	0.4688	0.3971
3.865	2.0333	864	0.5912	0.8711	0.8732	0.8711	0.8496
1.4285	3.0333	1152	0.3284	0.8867	0.8762	0.8867	0.8663
0.8195	4.0333	1440	0.1833	0.9375	0.9276	0.9375	0.9256
0.4318	5.0333	1728	0.0732	0.9688	0.9756	0.9688	0.9665
0.274	6.0333	2016	0.0440	0.9844	0.9885	0.9844	0.9837
0.2439	7.0333	2304	0.0661	0.9766	0.9792	0.9766	0.9762
0.1427	8.0333	2592	0.0970	0.9805	0.9734	0.9805	0.9752
0.137	9.0333	2880	0.0308	0.9961	0.9969	0.9961	0.9960
0.1534	10.0333	3168	0.0466	0.9883	0.9906	0.9883	0.9881
0.122	11.0333	3456	0.1174	0.9766	0.9784	0.9766	0.9763
0.0199	12.0333	3744	0.0210	0.9922	0.9922	0.9922	0.9922
0.0184	13.0333	4032	0.0467	0.9883	0.9917	0.9883	0.9877
0.1376	14.0333	4320	0.0025	1.0	1.0	1.0	1.0
0.0664	15.0333	4608	0.0323	0.9883	0.9906	0.9883	0.9881
0.1192	16.0333	4896	0.0807	0.9844	0.9885	0.9844	0.9837
0.0246	17.0333	5184	0.0387	0.9922	0.9930	0.9922	0.9921
0.0084	18.0333	5472	0.0339	0.9883	0.9906	0.9883	0.9881
0.0081	19.0333	5760	0.0353	0.9922	0.9938	0.9922	0.9921

Framework versions

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