VideoMAE_BdSLW60_FrameRate_Corrected_with_Augment_20_epoch_RQ

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: 0.6708
• Accuracy: 0.895
• Precision: 0.9032
• Recall: 0.895
• F1: 0.8843

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: 18560
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	Precision	Recall	F1
10.4448	0.0501	929	2.5241	0.3483	0.3333	0.3483	0.2892
2.3392	1.0501	1858	0.8054	0.7983	0.8006	0.7983	0.7746
0.8833	2.0501	2787	0.5077	0.8467	0.8835	0.8467	0.8400
0.3893	3.0501	3716	0.5973	0.8783	0.8885	0.8783	0.8613
0.4181	4.0501	4645	1.0255	0.795	0.8179	0.795	0.7654
0.144	5.0501	5574	0.4972	0.905	0.9065	0.905	0.8870
0.2381	6.0501	6503	1.1021	0.8017	0.8541	0.8017	0.7955
0.2042	7.0501	7432	0.9343	0.855	0.8919	0.855	0.8343
0.1844	8.0501	8361	0.4798	0.9083	0.9237	0.9083	0.9061
0.1416	9.0501	9290	0.5504	0.9	0.9332	0.9	0.8899
0.1182	10.0501	10219	0.3593	0.9317	0.9462	0.9317	0.9311
0.0255	11.0501	11148	0.5179	0.9	0.9307	0.9	0.8987
0.1122	12.0501	12077	0.5793	0.9017	0.9192	0.9017	0.8968
0.0681	13.0501	13006	0.6389	0.9133	0.9368	0.9133	0.9056
0.047	14.0501	13935	0.5920	0.9067	0.9284	0.9067	0.9054
0.0057	15.0501	14864	0.6708	0.895	0.9032	0.895	0.8843

Framework versions

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