convnextv2-base-22k-224-finetuned-eurosat-50

This model is a fine-tuned version of facebook/convnextv2-base-22k-224 on the imagefolder dataset. It achieves the following results on the evaluation set:

• Loss: 0.2875
• Accuracy: 0.9147

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-06
• train_batch_size: 16
• eval_batch_size: 16
• seed: 42
• gradient_accumulation_steps: 2
• total_train_batch_size: 32
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_ratio: 0.9
• num_epochs: 12

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
1.9019	1.0	122	1.9510	0.1727
1.7782	2.0	244	1.8239	0.3073
1.6214	3.0	366	1.6121	0.4913
1.3495	4.0	488	1.3064	0.6238
1.0994	5.0	610	1.0243	0.7163
0.8866	6.0	732	0.8165	0.7564
0.7282	7.0	854	0.6637	0.7996
0.6211	8.0	976	0.5623	0.8160
0.5114	9.0	1098	0.4681	0.8551
0.3835	10.0	1220	0.3917	0.8787
0.3543	11.0	1342	0.3122	0.9013
0.3534	12.0	1464	0.2875	0.9147

Framework versions

• Transformers 4.30.2
• Pytorch 2.0.1+cu118
• Datasets 2.13.1
• Tokenizers 0.13.3