videomae-base-finetuned-kinetics-allkisa-crop-background-0228-clip_duration

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

• Loss: 0.3168
• Accuracy: 0.9393

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: 8
• eval_batch_size: 8
• seed: 42
• optimizer: Use 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: 0.05
• training_steps: 26400

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
0.0014	0.0100	265	0.5277	0.7871
0.1062	1.0100	530	0.4787	0.8080
0.0267	2.0100	795	0.5039	0.8429
0.0023	3.0100	1060	1.0419	0.7033
0.0003	4.0100	1325	0.5443	0.8115
0.0002	5.0100	1590	1.1036	0.7644
0.0006	6.0100	1855	0.4425	0.8551
3.5553	7.0100	2120	0.4898	0.8586
0.0002	8.0100	2385	0.5075	0.8482
0.0	9.0100	2650	0.5837	0.8674
0.0001	10.0100	2915	0.9025	0.8255
0.0001	11.0100	3180	0.6071	0.8691
0.0	12.0100	3445	0.6740	0.8639
0.0	13.0100	3710	0.9123	0.8272
0.0	14.0100	3975	0.7934	0.8517
0.0	15.0100	4240	0.6899	0.8674
0.0	16.0100	4505	0.7532	0.8709
0.0	17.0100	4770	0.7553	0.8691
0.0002	18.0100	5035	0.7433	0.8691
0.0	19.0100	5300	0.6715	0.8813
0.0	20.0100	5565	0.7857	0.8656
0.0	21.0100	5830	0.6759	0.8918
0.0	22.0100	6095	0.6435	0.8901
0.0	23.0100	6360	0.6853	0.8866
0.0	24.0100	6625	0.6530	0.8831
0.0038	25.0100	6890	0.7277	0.8743
0.0001	26.0100	7155	0.7137	0.8778
0.0	27.0100	7420	0.7081	0.8796
0.0	28.0100	7685	0.6977	0.8778
0.0	29.0100	7950	0.7008	0.8831
0.0004	30.0100	8215	0.7023	0.8813
0.0	31.0100	8480	0.7248	0.8796

Framework versions

• Transformers 4.48.1
• Pytorch 2.5.1+cu121
• Datasets 3.2.0
• Tokenizers 0.21.0