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: 0.3585
• Accuracy: 0.9226

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: 2
• eval_batch_size: 2
• 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: 3750

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
2.0046	0.04	150	2.0806	0.3143
1.9549	1.04	300	1.4746	0.5429
0.6873	2.04	450	0.9489	0.6714
1.6042	3.04	600	0.6865	0.7571
0.2082	4.04	750	0.4017	0.8857
0.2805	5.04	900	0.9705	0.7714
0.0062	6.04	1050	0.4833	0.8571
0.2727	7.04	1200	0.8048	0.8714
0.1055	8.04	1350	0.0264	0.9857
0.234	9.04	1500	0.1460	0.9714
0.0015	10.04	1650	0.3039	0.9429
0.0012	11.04	1800	0.2351	0.9571
0.0009	12.04	1950	0.3080	0.9286
0.0009	13.04	2100	0.3477	0.9429
0.036	14.04	2250	0.2366	0.9571
0.0008	15.04	2400	0.4506	0.9
0.0037	16.04	2550	0.2327	0.9571
0.0007	17.04	2700	0.3480	0.9286
0.0007	18.04	2850	0.1762	0.9714
0.0006	19.04	3000	0.0991	0.9714
0.0006	20.04	3150	0.1551	0.9714
0.0006	21.04	3300	0.3023	0.9429
0.0006	22.04	3450	0.1543	0.9571
0.0006	23.04	3600	0.1025	0.9571
0.0006	24.04	3750	0.0876	0.9571

Framework versions

• Transformers 4.31.0
• Pytorch 2.0.1+cu117
• Datasets 2.14.4
• Tokenizers 0.13.3