hftest-MHmae

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: 5.4422
• Accuracy: 0.156

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: 16
• eval_batch_size: 16
• seed: 42
• gradient_accumulation_steps: 16
• total_train_batch_size: 256
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_ratio: 0.1
• num_epochs: 20

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
No log	0.97	19	6.0081	0.0026
No log	1.99	39	5.9817	0.0034
No log	2.96	58	5.9457	0.0084
No log	3.99	78	5.9030	0.0118
No log	4.96	97	5.8531	0.0204
No log	5.98	117	5.8012	0.0338
No log	6.95	136	5.7553	0.0458
No log	7.97	156	5.7005	0.0666
No log	9.0	176	5.6508	0.0842
No log	9.97	195	5.6129	0.1006
No log	10.99	215	5.5841	0.1108
No log	11.96	234	5.5503	0.1208
No log	12.98	254	5.5248	0.129
No log	13.96	273	5.5014	0.1372
No log	14.98	293	5.4761	0.1428
No log	16.0	313	5.4660	0.1484
No log	16.97	332	5.4574	0.1556
No log	17.99	352	5.4520	0.1546
No log	18.96	371	5.4458	0.1548
No log	19.42	380	5.4422	0.156

Framework versions

• Transformers 4.30.2
• Pytorch 2.1.0.dev20230627
• Datasets 2.13.1
• Tokenizers 0.13.3