README.md

metadata

tags:
  - generated_from_trainer
metrics:
  - accuracy
model-index:
  - name: videomae-large
    results: []

videomae-large

This model was trained from scratch on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 0.0993
• Accuracy: 0.9742

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: 32
• eval_batch_size: 32
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_ratio: 0.1
• training_steps: 300

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
2.3203	0.03	10	2.1994	0.1571
1.9795	1.03	20	1.7835	0.3429
1.0467	2.03	30	0.7311	0.6571
0.301	3.03	40	0.2195	0.9429
0.1061	4.03	50	0.1529	0.9143
0.0499	5.03	60	0.0826	0.9857
0.079	6.03	70	0.0534	0.9857
0.0487	7.03	80	0.0299	0.9857
0.0217	8.03	90	0.3283	0.9
0.0387	9.03	100	0.0268	0.9857
0.0252	10.03	110	0.0386	0.9857
0.0324	11.03	120	0.3067	0.9
0.0022	12.03	130	0.0131	1.0
0.0115	13.03	140	0.0889	0.9857
0.0225	14.03	150	0.0091	1.0
0.0012	15.03	160	0.0081	1.0
0.001	16.03	170	0.0103	1.0
0.0255	17.03	180	0.0113	1.0
0.0016	18.03	190	0.0252	0.9857
0.0039	19.03	200	0.0177	0.9857
0.0007	20.03	210	0.0017	1.0
0.0006	21.03	220	0.0013	1.0
0.0006	22.03	230	0.0012	1.0
0.0007	23.03	240	0.0011	1.0
0.0005	24.03	250	0.0011	1.0
0.0005	25.03	260	0.0011	1.0
0.0005	26.03	270	0.0011	1.0
0.0005	27.03	280	0.0011	1.0
0.0005	28.03	290	0.0011	1.0
0.0005	29.03	300	0.0011	1.0

Framework versions

• Transformers 4.37.2
• Pytorch 2.1.0+cu121
• Datasets 2.17.1
• Tokenizers 0.15.2