custom-cloud-model

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

• Loss: 0.2242
• Accuracy: 0.9506

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-05
• train_batch_size: 8
• eval_batch_size: 8
• seed: 42
• gradient_accumulation_steps: 4
• total_train_batch_size: 32
• optimizer: Use OptimizerNames.ADAMW_TORCH_FUSED with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: cosine
• lr_scheduler_warmup_ratio: 0.1
• num_epochs: 25
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
1.153	1.0	80	0.9964	0.7510
0.8046	2.0	160	0.4400	0.8477
0.7706	3.0	240	0.5513	0.8107
0.6695	4.0	320	0.2639	0.9136
0.4697	5.0	400	0.2199	0.9300
0.4564	6.0	480	0.1808	0.9321
0.4094	7.0	560	0.2471	0.9239
0.2445	8.0	640	0.1998	0.9424
0.3153	9.0	720	0.2600	0.9136
0.1994	10.0	800	0.3160	0.9259
0.2205	11.0	880	0.2486	0.9239
0.2312	12.0	960	0.2131	0.9486
0.2382	13.0	1040	0.2487	0.9239
0.1158	14.0	1120	0.2153	0.9506
0.1	15.0	1200	0.2271	0.9486
0.0916	16.0	1280	0.2626	0.9259
0.1021	17.0	1360	0.2130	0.9403
0.0991	18.0	1440	0.2240	0.9444
0.0526	19.0	1520	0.2353	0.9506
0.0461	20.0	1600	0.2137	0.9506
0.0832	21.0	1680	0.2215	0.9444
0.0708	22.0	1760	0.2114	0.9506
0.1027	23.0	1840	0.2104	0.9527
0.0292	24.0	1920	0.2346	0.9547
0.033	25.0	2000	0.2242	0.9506

Framework versions

• Transformers 4.57.1
• Pytorch 2.9.0+cu126
• Datasets 4.0.0
• Tokenizers 0.22.1