swinv2-tiny-patch4-window8-256-dmae-humeda-DAV39

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: 0.9091
• Accuracy: 0.75

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: 32
• eval_batch_size: 32
• seed: 42
• gradient_accumulation_steps: 4
• total_train_batch_size: 128
• 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: 30

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
No log	1.0	4	1.3890	0.4318
No log	2.0	8	1.2016	0.5682
4.424	3.0	12	1.0279	0.6364
4.424	4.0	16	0.9227	0.6477
4.424	5.0	20	0.9294	0.6818
2.6292	6.0	24	0.9936	0.625
2.6292	7.0	28	0.9816	0.6705
2.6292	8.0	32	0.9091	0.75
1.5117	9.0	36	0.8995	0.7159
1.5117	10.0	40	1.0646	0.7045
1.5117	11.0	44	1.1826	0.6705
0.7053	12.0	48	1.2360	0.6932
0.7053	13.0	52	1.4215	0.6818
0.7053	14.0	56	1.3694	0.7045
0.2599	15.0	60	1.2998	0.7159
0.2599	16.0	64	1.3703	0.7159
0.2599	17.0	68	1.4626	0.7159
0.1282	18.0	72	1.5181	0.7159
0.1282	19.0	76	1.5692	0.7045
0.1282	20.0	80	1.5965	0.6932
0.0816	21.0	84	1.6018	0.7045
0.0816	22.0	88	1.6046	0.7045
0.0816	22.6154	90	1.6049	0.7045

Framework versions

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