videomae-large-kissing_14-01-2024

This model is a fine-tuned version of MCG-NJU/videomae-large on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 0.3655
• Accuracy: 0.9479
• Precision: 0.9547
• Recall: 0.9405

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-07
• 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: 18165

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	Precision	Recall
0.6377	0.07	1212	0.6211	0.6645	0.6755	0.6331
0.586	1.07	2424	0.4979	0.7835	0.8057	0.7471
0.3675	2.07	3636	0.3910	0.8983	0.9335	0.8579
0.8145	3.07	4848	0.3776	0.9207	0.9426	0.8959
0.6408	4.07	6060	0.3674	0.9322	0.9470	0.9157
0.01	5.07	7272	0.3630	0.9298	0.9422	0.9157
0.0274	6.07	8484	0.3808	0.9289	0.9233	0.9355
0.0002	7.07	9696	0.3566	0.9397	0.9508	0.9273
0.0058	8.07	10908	0.3609	0.9446	0.9622	0.9256
0.1551	9.07	12120	0.3757	0.9413	0.9465	0.9355
0.1784	10.07	13332	0.3410	0.9496	0.9579	0.9405
0.0011	11.07	14544	0.3707	0.9455	0.9455	0.9455
0.0001	12.07	15756	0.3719	0.9479	0.9547	0.9405
0.0307	13.07	16968	0.3657	0.9463	0.9530	0.9388
0.0002	14.07	18165	0.3655	0.9479	0.9547	0.9405

Framework versions

• Transformers 4.35.2
• Pytorch 2.1.0+cu121
• Datasets 2.16.1
• Tokenizers 0.15.0