videomae-base-finetuned-hmdb51_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: 0.3945
• Accuracy: 0.8798

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: 32
• eval_batch_size: 32
• seed: 42
• 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: 380

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
1.8738	0.1026	39	1.6265	0.4299
1.1232	1.1026	78	0.9534	0.7430
0.86	2.1026	117	0.7309	0.7897
0.5001	3.1026	156	0.5999	0.8131
0.4402	4.1026	195	0.5475	0.8178
0.3291	5.1026	234	0.5903	0.8505
0.2643	6.1026	273	0.5727	0.8271
0.1928	7.1026	312	0.4194	0.8972
0.1749	8.1026	351	0.4278	0.8832
0.1588	9.0763	380	0.4737	0.8832

Framework versions

• Transformers 4.51.3
• Pytorch 2.6.0+cu124
• Datasets 3.5.1
• Tokenizers 0.21.1