videomae-base-finetuned-kinetics-allkisa-crop-background-0311-clip_duration-abnormal12

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: 0.1913
• Accuracy: 0.9521

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: 29600

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
0.7497	0.01	296	0.3402	0.8788
0.1554	1.01	592	0.4850	0.8481
0.1181	2.01	888	0.8904	0.7851
0.6108	3.01	1184	0.4483	0.8772
0.0009	4.01	1480	0.4471	0.8611
0.3527	5.01	1776	0.6108	0.8336
0.0845	6.01	2072	0.6113	0.8498
0.5731	7.01	2368	0.6838	0.8255
0.0085	8.01	2664	0.3924	0.9079
0.0022	9.01	2960	0.5572	0.8740
0.0011	10.01	3256	0.5516	0.8578
0.0007	11.01	3552	0.4634	0.8869
0.0056	12.01	3848	0.5271	0.8821
0.0056	13.01	4144	0.6203	0.8675
0.0002	14.01	4440	0.5221	0.8934
0.0001	15.01	4736	0.5882	0.8885
0.0001	16.01	5032	0.5987	0.8853
0.0003	17.01	5328	0.5436	0.9031
0.0082	18.01	5624	0.5575	0.8966

Framework versions

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