hurricane_model

This model is a fine-tuned version of google/vit-base-patch16-224-in21k on the jonathan-roberts1/Satellite-Images-of-Hurricane-Damage dataset. It achieves the following results on the evaluation set:

• Loss: 0.0224
• Model Preparation Time: 0.0051
• Accuracy: 0.9948

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: 16
• eval_batch_size: 8
• seed: 42
• optimizer: Use adamw_torch with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: linear
• num_epochs: 4
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Model Preparation Time	Accuracy
0.1118	0.3195	100	0.1486	0.0051	0.9476
0.1112	0.6390	200	0.0701	0.0051	0.9752
0.0694	0.9585	300	0.0608	0.0051	0.9808
0.0048	1.2780	400	0.0917	0.0051	0.9744
0.036	1.5974	500	0.0552	0.0051	0.9836
0.0594	1.9169	600	0.0547	0.0051	0.9808
0.0115	2.2364	700	0.0627	0.0051	0.9844
0.0016	2.5559	800	0.0296	0.0051	0.9936
0.004	2.8754	900	0.0325	0.0051	0.9916
0.0009	3.1949	1000	0.0224	0.0051	0.9948
0.0008	3.5144	1100	0.0270	0.0051	0.9936
0.0008	3.8339	1200	0.0256	0.0051	0.994

Framework versions

• Transformers 4.46.3
• Pytorch 2.5.1+cu121
• Datasets 3.2.0
• Tokenizers 0.20.3