jialicheng/unlearn_cifar100_swin-base_random_label_2_87

87

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: 0.4135
• Accuracy: 0.9046
• Dt Accuracy: 0.9046
• Df Accuracy: 0.027
• Unlearn Overall Accuracy: 0.9787
• 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: 0.0002
• train_batch_size: 128
• eval_batch_size: 256
• seed: 87
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 20

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	Overall Accuracy	Unlearn Overall Accuracy	Time
No log	1.0	391	0.4750	0.878	0.2313	0.2313	None
1.1674	2.0	782	0.4494	0.878	0.2315	0.2315	None
1.0003	3.0	1173	0.4268	0.882	0.2255	0.2255	None
0.8869	4.0	1564	0.4526	0.841	0.2865	0.2865	None
0.8869	5.0	1955	0.4342	0.818	0.3194	0.3194	None
0.7869	6.0	2346	0.4292	0.745	0.4148	0.4148	None
0.7179	7.0	2737	0.4242	0.656	0.5171	0.5171	None
0.6474	8.0	3128	0.4275	0.537	0.6330	0.6330	None
0.5895	9.0	3519	0.4554	0.468	0.6915	0.6915	None
0.5895	10.0	3910	0.4530	0.372	0.7668	0.7668	None
0.5329	11.0	4301	0.4584	0.302	0.8129	0.8129	None
0.4855	12.0	4692	0.4445	0.252	0.8482	0.8482	None
0.4446	13.0	5083	0.4372	0.168	0.8984	0.8984	None
0.4446	14.0	5474	0.4565	0.141	0.9134	0.9134	None
0.4036	15.0	5865	0.4320	0.113	0.9313	0.9313	None
0.3756	16.0	6256	0.4295	0.053	0.9631	0.9631	None
0.3501	17.0	6647	0.4035	0.058	0.9625	0.9625	None
0.3236	18.0	7038	0.4261	0.038	0.9702	0.9702	None
0.3236	19.0	7429	0.4121	0.031	0.9762	0.9762	None
0.2996	20.0	7820	0.4135	0.027	0.9787	0.9787	None

Framework versions

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