videomae-base-finetuned-kinetics-0328_final_253259

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: 0.5132
• Accuracy: 0.8633

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: 72700

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
0.1593	0.0100	728	0.7166	0.7523
0.0076	1.0100	1456	0.3380	0.9012
0.0034	2.0100	2184	0.3450	0.9037
0.0029	3.0100	2912	0.6705	0.8041
0.0018	4.0100	3640	0.8537	0.8140
0.1048	5.0100	4368	0.4160	0.8872
0.0186	6.0100	5096	0.7927	0.8107
0.0017	7.0100	5824	0.7084	0.8123
0.0003	8.0100	6552	0.7571	0.8560
0.003	9.0100	7280	0.4314	0.8988
0.0004	10.0100	8008	0.9204	0.8165
0.0002	11.0100	8736	1.2702	0.7901
0.0002	12.0100	9464	0.5825	0.8831
0.0004	13.0100	10192	0.9033	0.8288
0.0002	14.0100	10920	0.9259	0.8527
0.0051	15.0100	11648	1.0172	0.8337
0.0002	16.0100	12376	1.0718	0.8214
0.0	17.0100	13104	1.0214	0.8444

Framework versions

• Transformers 4.48.1
• Pytorch 2.5.1+cu121
• Datasets 3.2.0
• Tokenizers 0.21.0