videomae-base-finetuned-ucf101-subset2

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: 0.4611
• Accuracy: 0.9286

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: 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: 1200

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
0.0018	0.06	75	0.2489	0.9
0.0033	1.06	150	0.6663	0.8429
0.7225	2.06	225	1.5428	0.7143
0.9709	3.06	300	0.5602	0.8571
0.0012	4.06	375	0.5840	0.8857
0.0471	5.06	450	0.8610	0.8429
0.0008	6.06	525	0.4117	0.9
0.0007	7.06	600	0.4993	0.9
0.0005	8.06	675	0.6722	0.8571
0.0252	9.06	750	0.4827	0.8714
0.0005	10.06	825	0.5150	0.9286
0.0005	11.06	900	0.4033	0.9286
0.0005	12.06	975	0.4546	0.9286
0.0004	13.06	1050	0.4545	0.9286
0.0004	14.06	1125	0.4596	0.9286
0.0005	15.06	1200	0.4611	0.9286

Framework versions

• Transformers 4.30.2
• Pytorch 1.8.0
• Datasets 2.13.2
• Tokenizers 0.13.3