swinv2-tiny-patch4-window8-256-dmae-humeda-DAV57

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: 0.8727
• Accuracy: 0.7308

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: 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_with_restarts
• lr_scheduler_warmup_ratio: 0.1
• num_epochs: 45
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
No log	0.8421	4	1.5807	0.2885
No log	1.8421	8	1.4549	0.4423
1.6981	2.8421	12	1.3224	0.4231
1.6981	3.8421	16	1.2079	0.5
1.6981	4.8421	20	1.0541	0.5769
1.2915	5.8421	24	0.9398	0.6346
1.2915	6.8421	28	0.9887	0.5577
1.2915	7.8421	32	0.8991	0.6538
0.8599	8.8421	36	0.9379	0.5577
0.8599	9.8421	40	0.8260	0.6923
0.8599	10.8421	44	0.9418	0.6731
0.6803	11.8421	48	0.9368	0.5769
0.6803	12.8421	52	0.9148	0.5962
0.6803	13.8421	56	0.9135	0.6346
0.5562	14.8421	60	0.8477	0.6731
0.5562	15.8421	64	0.8730	0.5962
0.5562	16.8421	68	0.8420	0.6923
0.4696	17.8421	72	0.9168	0.5962
0.4696	18.8421	76	0.9373	0.6538
0.4696	19.8421	80	0.8634	0.6538
0.3975	20.8421	84	0.8695	0.6538
0.3975	21.8421	88	0.8958	0.6923
0.3975	22.8421	92	0.8914	0.6731
0.3185	23.8421	96	0.8727	0.7308
0.3185	24.8421	100	0.9820	0.6923
0.3185	25.8421	104	0.9263	0.6923
0.2758	26.8421	108	1.0548	0.5962
0.2758	27.8421	112	0.9833	0.6731
0.2758	28.8421	116	0.9492	0.6923
0.2667	29.8421	120	0.9466	0.6346
0.2667	30.8421	124	0.9828	0.6923
0.2667	31.8421	128	1.1056	0.6923
0.2396	32.8421	132	1.0083	0.6731
0.2396	33.8421	136	1.0040	0.6923
0.2396	34.8421	140	1.0727	0.6731
0.2173	35.8421	144	1.0953	0.6923
0.2173	36.8421	148	1.0802	0.6538
0.2173	37.8421	152	1.0446	0.6923
0.2313	38.8421	156	1.0331	0.7115
0.2313	39.8421	160	1.0334	0.6923
0.2313	40.8421	164	1.0364	0.6923
0.2129	41.8421	168	1.0413	0.6731
0.2129	42.8421	172	1.0407	0.6731
0.2129	43.8421	176	1.0405	0.6731
0.2026	44.8421	180	1.0401	0.6731

Framework versions

• Transformers 4.48.2
• Pytorch 2.5.1+cu124
• Datasets 3.2.0
• Tokenizers 0.21.0