videomae-base-finetuned-midv-holo-500

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: 2.8322
• Accuracy: 0.4264

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: 4
• eval_batch_size: 4
• seed: 42
• optimizer: Use OptimizerNames.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: 0.1
• training_steps: 1925

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
1.2647	0.0405	78	1.2009	0.5736
1.3169	1.0405	156	1.2652	0.5078
1.3004	2.0405	234	1.2036	0.5736
1.1562	3.0405	312	1.1888	0.5736
1.265	4.0405	390	1.2808	0.4922
1.1946	5.0405	468	1.3394	0.3295
1.1222	6.0405	546	1.3204	0.5
1.0244	7.0405	624	1.3697	0.3643
1.1604	8.0405	702	1.3813	0.4535
0.8901	9.0405	780	1.4395	0.4225
0.6032	10.0405	858	1.6609	0.4535
0.6889	11.0405	936	1.7041	0.3605
0.5777	12.0405	1014	1.9075	0.3915
0.4317	13.0405	1092	1.6528	0.4690
0.5087	14.0405	1170	1.6126	0.4845
0.1257	15.0405	1248	2.0637	0.4147
0.3774	16.0405	1326	2.5347	0.3721
0.1259	17.0405	1404	2.6514	0.3760
0.2158	18.0405	1482	2.4290	0.4806
0.0505	19.0405	1560	2.6998	0.4225
0.0563	20.0405	1638	2.6474	0.4225
0.0671	21.0405	1716	2.6487	0.4496
0.1847	22.0405	1794	2.7012	0.4496
0.0041	23.0405	1872	2.7938	0.4341
0.0079	24.0275	1925	2.8322	0.4264

Framework versions

• Transformers 4.47.1
• Pytorch 2.5.1+cu118
• Datasets 3.2.0
• Tokenizers 0.21.0