swinv2-tiny-patch4-window8-256-dmae-humeda-DAV43

This model is a fine-tuned version of microsoft/swinv2-tiny-patch4-window8-256 on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 1.0777
• Accuracy: 0.6374

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: 2e-05
• train_batch_size: 32
• eval_batch_size: 32
• seed: 42
• gradient_accumulation_steps: 8
• total_train_batch_size: 256
• optimizer: Use OptimizerNames.ADAMW_TORCH 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: 42

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
No log	1.0	2	1.5998	0.1319
No log	2.0	4	1.4960	0.4176
No log	3.0	6	1.3659	0.4396
No log	4.0	8	1.3074	0.4286
No log	5.0	10	1.2850	0.4286
9.4768	6.0	12	1.2592	0.4396
9.4768	7.0	14	1.2446	0.4505
9.4768	8.0	16	1.2326	0.5495
9.4768	9.0	18	1.2084	0.5714
9.4768	10.0	20	1.1779	0.5604
9.4768	11.0	22	1.1483	0.5604
7.6512	12.0	24	1.1232	0.5824
7.6512	13.0	26	1.1042	0.6044
7.6512	14.0	28	1.0891	0.6154
7.6512	15.0	30	1.0777	0.6374
7.6512	16.0	32	1.0686	0.6264
7.6512	17.0	34	1.0623	0.6264
6.5906	18.0	36	1.0582	0.6264
6.5906	19.0	38	1.0559	0.6154
6.5906	20.0	40	1.0550	0.6154
6.5906	21.0	42	1.0548	0.6154

Framework versions

• Transformers 4.47.1
• Pytorch 2.5.1+cu124
• Datasets 3.2.0
• Tokenizers 0.21.0