videomae-base-finetuned-ucf101-subset

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: 0.1034
• Accuracy: 0.9429

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: 64
• eval_batch_size: 64
• 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: 148

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
No log	0.03	5	2.2559	0.1143
2.2949	1.03	10	2.1762	0.1429
2.2949	2.03	15	2.0958	0.2286
2.0288	3.03	20	1.8294	0.3
2.0288	4.03	25	1.4324	0.5857
1.3182	5.03	30	1.1143	0.6714
1.3182	6.03	35	0.8853	0.7571
0.6119	7.03	40	0.6069	0.8143
0.6119	8.03	45	0.5380	0.8
0.3056	9.03	50	0.3780	0.8571
0.3056	10.03	55	0.3159	0.9
0.1774	11.03	60	0.3172	0.9143
0.1774	12.03	65	0.2780	0.8857
0.1209	13.03	70	0.3128	0.8857
0.1209	14.03	75	0.3360	0.8857
0.0805	15.03	80	0.2165	0.9143
0.0805	16.03	85	0.2362	0.9429
0.0382	17.03	90	0.2397	0.9143
0.0382	18.03	95	0.1855	0.9
0.0268	19.03	100	0.2475	0.9
0.0268	20.03	105	0.2548	0.9143
0.0227	21.03	110	0.1521	0.9571
0.0227	22.03	115	0.1162	0.9286
0.0168	23.03	120	0.1924	0.9429
0.0168	24.03	125	0.0873	0.9571
0.0143	25.03	130	0.1972	0.9429
0.0143	26.03	135	0.1505	0.9429
0.0168	27.03	140	0.1022	0.9571
0.0168	28.03	145	0.1014	0.9429
0.0168	29.02	148	0.1034	0.9429

Framework versions

• Transformers 4.30.2
• Pytorch 1.13.0+cu117
• Datasets 2.13.2
• Tokenizers 0.13.3