videomae-base-finetuned-deception-dataset

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: 1.8677
• Accuracy: 0.5679

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: 13
• eval_batch_size: 13
• seed: 42
• gradient_accumulation_steps: 2
• total_train_batch_size: 26
• optimizer: Use OptimizerNames.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: 368

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
0.6819	1.0	24	0.6935	0.4444
0.6136	2.0	48	0.9791	0.5185
0.3068	3.0	72	1.5908	0.5679
0.2164	4.0	96	0.9361	0.5926
0.2	5.0	120	1.6885	0.5309
0.3087	6.0	144	0.9806	0.7284
0.1689	7.0	168	1.3332	0.5556
0.1326	8.0	192	1.2937	0.5802
0.1132	9.0	216	1.7412	0.5432
0.0967	10.0	240	1.8026	0.5556
0.1155	11.0	264	2.0465	0.5556
0.0826	12.0	288	2.1887	0.5556
0.1027	13.0	312	2.1518	0.5432
0.0714	14.0	336	2.1663	0.5432
0.0654	15.0	360	1.8426	0.5679
0.0654	15.3404	368	1.8677	0.5679

Framework versions

• Transformers 4.51.3
• Pytorch 2.1.0+cu121
• Datasets 3.5.0
• Tokenizers 0.21.1