ALL_MVCROP_ori16F-8B16F-GACWDlr

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: 1.5798
• Accuracy: 0.6766

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: 1e-05
• train_batch_size: 8
• eval_batch_size: 8
• seed: 42
• gradient_accumulation_steps: 2
• total_train_batch_size: 16
• 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: 1920

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
0.6954	0.025	48	0.6919	0.4878
0.696	1.025	96	0.6908	0.5244
0.6338	2.025	144	0.6795	0.5671
0.5972	3.025	192	0.6620	0.5793
0.5861	4.025	240	0.8272	0.5366
0.3748	5.025	288	0.6881	0.6585
0.3576	6.025	336	0.7236	0.6585
0.366	7.025	384	0.8092	0.6707
0.1977	8.025	432	0.8176	0.6585
0.1381	9.025	480	1.0842	0.6890
0.1265	10.025	528	1.5289	0.5793
0.0772	11.025	576	1.4717	0.6341
0.0116	12.025	624	1.4711	0.7134
0.0989	13.025	672	2.1604	0.5976
0.0676	14.025	720	1.6358	0.6951
0.0032	15.025	768	1.6290	0.7012
0.0385	16.025	816	1.8245	0.6646

Framework versions

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