human_action_recognition_model

This model is a fine-tuned version of google/vit-base-patch16-224-in21k on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 7.8069
• Accuracy: 0.0659

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: 0.0002
• train_batch_size: 8
• eval_batch_size: 8
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 4

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
1.3102	0.3175	500	3.5439	0.0761
0.9861	0.6349	1000	4.1324	0.065
0.8791	0.9524	1500	4.6708	0.0752
0.5281	1.2698	2000	5.0605	0.0980
0.4598	1.5873	2500	6.1627	0.0437
0.4733	1.9048	3000	5.6746	0.0754
0.2844	2.2222	3500	6.5390	0.0746
0.1697	2.5397	4000	6.9396	0.0537
0.1697	2.8571	4500	7.1644	0.0672
0.1013	3.1746	5000	7.4083	0.0619
0.0556	3.4921	5500	7.4283	0.0694
0.0338	3.8095	6000	7.8069	0.0659

Framework versions

• Transformers 4.44.2
• Pytorch 2.4.0+cu121
• Datasets 3.0.0
• Tokenizers 0.19.1