swinv2-tiny-patch4-window8-256-dmae-humeda-DAV24

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: 0.8851
• Accuracy: 0.7059

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: 3e-05
• 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: 0.2
• num_epochs: 25

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
6.5614	1.0	17	1.6096	0.3176
6.1279	2.0	34	1.5651	0.3176
5.5089	3.0	51	1.3188	0.5529
4.453	4.0	68	1.0195	0.6353
3.3808	5.0	85	0.9741	0.5882
2.7707	6.0	102	0.8365	0.6353
2.3091	7.0	119	0.7725	0.6588
1.9831	8.0	136	0.8312	0.6588
1.8284	9.0	153	0.8473	0.7059
1.511	10.0	170	0.7539	0.7176
1.2827	11.0	187	0.8067	0.7176
1.2072	12.0	204	0.7927	0.7176
1.2069	13.0	221	0.8184	0.6824
0.9242	14.0	238	0.8548	0.7059
0.9772	15.0	255	0.8374	0.7294
0.8412	16.0	272	0.8340	0.7176
0.8921	17.0	289	0.8729	0.6941
0.7975	18.0	306	0.9115	0.7059
0.8107	19.0	323	0.8830	0.6941
0.7131	20.0	340	0.9049	0.6941
0.6777	21.0	357	0.8895	0.7059
0.6557	22.0	374	0.8831	0.7059
0.6555	23.0	391	0.8846	0.7059
0.7766	23.5455	400	0.8851	0.7059

Framework versions

• Transformers 4.47.1
• Pytorch 2.5.1+cu121
• Datasets 3.2.0
• Tokenizers 0.21.0