jialicheng/unlearn_cifar100_swin-base_random_label_4_42

42

This model is a fine-tuned version of microsoft/swin-base-patch4-window7-224 on the cifar100 dataset. It achieves the following results on the evaluation set:

• Loss: 0.3410
• Accuracy: 0.9174
• Dt Accuracy: 0.9174
• Df Accuracy: 0.381
• Unlearn Overall Accuracy: 0.4160
• Unlearn Time: None

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: 0.0001
• train_batch_size: 128
• eval_batch_size: 256
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 10

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	Overall Accuracy	Unlearn Overall Accuracy	Time
0.9711	1.0	391	0.4050	0.917	0.1537	0.1537	None
0.8672	2.0	782	0.3663	0.9175	0.1537	0.1537	None
0.7969	3.0	1173	0.3551	0.885	0.1763	0.1763	None
0.7222	4.0	1564	0.3826	0.798	0.2304	0.2304	None
0.6411	5.0	1955	0.3555	0.745	0.2605	0.2605	None
0.5803	6.0	2346	0.3580	0.6495	0.3088	0.3088	None
0.5138	7.0	2737	0.3436	0.552	0.3520	0.3520	None
0.4927	8.0	3128	0.3561	0.433	0.3972	0.3972	None
0.4468	9.0	3519	0.3412	0.409	0.4068	0.4068	None
0.4187	10.0	3910	0.3410	0.381	0.4160	0.4160	None

Framework versions

• Transformers 4.39.3
• Pytorch 2.2.2+cu118
• Datasets 2.18.0
• Tokenizers 0.15.2