---

license: apache-2.0 base_model: microsoft/swin-tiny-patch4-window7-224 tags:

• generated_from_trainer datasets:
• imagefolder metrics:
• accuracy model-index:
• name: cards_bottom_right_swin-tiny-patch4-window7-224-finetuned-v2 results:
  • task: name: Image Classification type: image-classification dataset: name: imagefolder type: imagefolder config: default split: test args: default metrics:
    • name: Accuracy type: accuracy value: 0.6078575555438837

---

cards_bottom_right_swin-tiny-patch4-window7-224-finetuned-v2

该模型是基于 microsoft/swin-tiny-patch4-window7-224 在 imagefolder 数据集上微调的版本。 它在评估集上取得了以下结果：

• Loss: 0.9317
• Accuracy: 0.6079

模型描述

更多信息待补充

预期用途与限制

更多信息待补充

训练与评估数据

更多信息待补充

训练过程

训练超参数

在训练过程中使用了以下超参数：

• learning_rate: 5e-05
• train_batch_size: 32
• eval_batch_size: 32
• seed: 42
• gradient_accumulation_steps: 4
• total_train_batch_size: 128
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_ratio: 0.1
• num_epochs: 30

训练结果

Training Loss	Epoch	Step	Validation Loss	Accuracy
1.4965	1.0	1338	1.3516	0.4156
1.4486	2.0	2677	1.1784	0.4938
1.4384	3.0	4015	1.1050	0.5223
1.4538	4.0	5354	1.0751	0.5433
1.3928	5.0	6692	1.0604	0.5440
1.4148	6.0	8031	1.0459	0.5523
1.3921	7.0	9369	1.0464	0.5501
1.3812	8.0	10708	1.0461	0.5491
1.3494	9.0	12046	1.0445	0.5486
1.3555	10.0	13385	0.9973	0.5693
1.3303	11.0	14723	0.9952	0.5719
1.3575	12.0	16062	1.0317	0.5574
1.3129	13.0	17400	0.9851	0.5813
1.3439	14.0	18739	1.0510	0.5523
1.3371	15.0	20077	0.9820	0.5795
1.2835	16.0	21416	0.9886	0.5738
1.3002	17.0	22754	0.9685	0.5869
1.289	18.0	24093	0.9519	0.5941
1.3007	19.0	25431	0.9855	0.5800
1.2927	20.0	26770	0.9499	0.5925
1.2985	21.0	28108	0.9669	0.5854
1.2957	22.0	29447	0.9551	0.5903
1.2579	23.0	30785	0.9300	0.6053
1.2475	24.0	32124	0.9296	0.6049
1.2227	25.0	33462	0.9317	0.6079
1.2069	26.0	34801	0.9609	0.5887
1.2156	27.0	36139	0.9297	0.6052
1.25	28.0	37478	0.9300	0.6062
1.2394	29.0	38816	0.9238	0.6071
1.209	29.99	40140	0.9284	0.6064

框架版本

• Transformers 4.37.2
• Pytorch 2.0.1+cu117
• Datasets 2.17.0
• Tokenizers 0.15.2