Real_RGBCROP_Aug16F-8B16F5e6-poly

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: 1.0455
• Accuracy: 0.7516

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-06
• train_batch_size: 8
• eval_batch_size: 8
• 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: polynomial
• lr_scheduler_warmup_ratio: 0.1
• training_steps: 2112

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
0.5535	0.0838	177	0.6608	0.6
0.2099	1.0838	354	0.6066	0.6967
0.1234	2.0838	531	0.7195	0.71
0.0375	3.0838	708	0.9196	0.7233
0.0023	4.0838	885	1.1401	0.72
0.0052	5.0838	1062	1.1748	0.73
0.002	6.0838	1239	1.2629	0.7133
0.0007	7.0838	1416	1.3239	0.7167
0.0005	8.0838	1593	1.3644	0.7233
0.0007	9.0838	1770	1.3876	0.7167

Framework versions

• Transformers 4.51.3
• Pytorch 2.6.0+cu124
• Datasets 3.6.0
• Tokenizers 0.21.1