videomae-base-finetuned-kinetics-finetuned-data-no-yolo-colab

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: 0.3187
• Accuracy: 0.9237

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: 9e-05
• train_batch_size: 4
• eval_batch_size: 4
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_ratio: 0.1
• training_steps: 2057

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
0.5766	0.09	188	0.7504	0.7956
0.4124	1.09	376	0.4797	0.8638
0.104	2.09	564	0.5037	0.8583
0.0652	3.09	752	0.5683	0.8556
0.0821	4.09	940	0.4314	0.8992
0.0015	5.09	1128	0.5824	0.8883
0.0024	6.09	1316	0.4310	0.8883
0.0007	7.09	1504	0.4375	0.8992
0.0007	8.09	1692	0.3199	0.9183
0.0006	9.09	1880	0.3189	0.9210
0.0006	10.09	2057	0.3187	0.9237

Framework versions

• Transformers 4.39.3
• Pytorch 2.2.2+cu121
• Datasets 2.18.0
• Tokenizers 0.15.2