videomae-base-finetuned-kinetics-0403_final_only5sec_retry

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: 0.8258
• Accuracy: 0.7900

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: 62100

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
0.0811	0.0100	622	0.3101	0.8648
0.0068	1.0100	1244	0.5977	0.8141
0.0011	2.0100	1866	0.7745	0.8203
0.0276	3.0100	2488	0.6727	0.8541
0.6154	4.0100	3110	0.6580	0.8399
0.1148	5.0100	3732	0.7695	0.8274
0.0053	6.0100	4354	0.8514	0.7945
0.0007	7.0100	4976	0.7208	0.8238
0.0014	8.0100	5598	1.3792	0.7375
0.0027	9.0100	6220	0.3791	0.8959
1.163	10.0100	6842	1.3893	0.7456
0.0178	11.0100	7464	1.1741	0.7820
0.009	12.0100	8086	0.6941	0.8470
0.004	13.0100	8708	0.9689	0.8274
1.0064	14.0100	9330	0.7809	0.8230
0.0003	15.0100	9952	1.1843	0.7918
0.0012	16.0100	10574	0.6599	0.8594
0.0395	17.0100	11196	0.6761	0.8799
0.0005	18.0100	11818	1.0149	0.8158
0.7313	19.0100	12440	1.3613	0.7811
0.0004	20.0100	13062	0.9187	0.8390
1.638	21.0100	13684	0.8302	0.8692
0.0002	22.0100	14306	0.9276	0.8568
0.0011	23.0100	14928	1.0031	0.8568
0.0002	24.0100	15550	0.9453	0.8568

Framework versions

• Transformers 4.48.0
• Pytorch 2.5.1+cu124
• Datasets 3.2.0
• Tokenizers 0.21.0