videomae-base-finetuned-kinetics-0313-clip_duration-abnormal02

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

• Loss: 0.2367
• Accuracy: 0.9553

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: 8
• eval_batch_size: 8
• seed: 42
• 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: 0.05
• training_steps: 35000

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
0.0666	0.01	350	0.6833	0.7785
0.2426	1.01	700	0.3021	0.8939
0.2266	2.01	1050	0.3887	0.8621
0.1474	3.01	1400	0.3437	0.8660
0.0026	4.01	1750	0.4075	0.8727
0.0385	5.01	2100	0.3817	0.8979
0.0054	6.01	2450	0.5044	0.8793
0.0002	7.01	2800	0.6241	0.8714
0.0155	8.01	3150	0.5603	0.8820
0.31	9.01	3500	0.4432	0.8992
0.0004	10.01	3850	0.4870	0.8899
0.0005	11.01	4200	0.5764	0.8979
0.0321	12.01	4550	0.4764	0.9032
0.0001	13.01	4900	0.5965	0.8966
0.0004	14.01	5250	0.5833	0.8939
0.0001	15.01	5600	0.6168	0.8833
0.0085	16.01	5950	0.5132	0.9072
0.0	17.01	6300	0.5239	0.9045
0.1187	18.01	6650	0.5492	0.8926
0.0001	19.01	7000	0.6863	0.8793
0.0001	20.01	7350	0.7093	0.8859
0.0	21.01	7700	0.5853	0.8912
0.0	22.01	8050	0.5131	0.9045
0.0	23.01	8400	0.6446	0.8912
0.0	24.01	8750	0.5270	0.9045
0.0	25.01	9100	0.5410	0.8992
0.0266	26.01	9450	0.6122	0.8899

Framework versions

• Transformers 4.48.0
• Pytorch 2.5.1+cu124
• Datasets 3.2.0
• Tokenizers 0.21.0