swin-tiny-patch4-window7-224-finetuned-tekno24-highdata-90-3rd

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

• Loss: 0.9108
• Accuracy: 0.6221
• F1: 0.6116
• Precision: 0.6170
• Recall: 0.6221

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: 16
• eval_batch_size: 16
• seed: 42
• gradient_accumulation_steps: 8
• total_train_batch_size: 128
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: cosine_with_restarts
• lr_scheduler_warmup_ratio: 0.1
• num_epochs: 30
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	F1	Precision	Recall
1.3622	0.9786	40	1.3459	0.3687	0.2453	0.2574	0.3687
1.335	1.9817	81	1.3295	0.3779	0.3635	0.3592	0.3779
1.2404	2.9847	122	1.1256	0.5023	0.4893	0.5048	0.5023
1.2113	3.9878	163	1.1081	0.5346	0.4918	0.5409	0.5346
1.1617	4.9908	204	1.0667	0.5300	0.4938	0.5204	0.5300
1.1758	5.9939	245	1.1505	0.4747	0.4713	0.5074	0.4747
1.1618	6.9969	286	1.1316	0.4931	0.4779	0.4950	0.4931
1.1748	8.0	327	1.0681	0.5161	0.4827	0.5256	0.5161
1.1421	8.9786	367	0.9743	0.5714	0.5445	0.5488	0.5714
1.1565	9.9817	408	0.9705	0.5622	0.5142	0.5429	0.5622
1.1297	10.9847	449	0.9879	0.5530	0.5365	0.5343	0.5530
1.1249	11.9878	490	0.9852	0.5760	0.5401	0.6055	0.5760
1.1289	12.9908	531	0.9555	0.5714	0.5363	0.5409	0.5714
1.1102	13.9939	572	0.9438	0.5991	0.5795	0.6033	0.5991
1.1011	14.9969	613	0.9492	0.5991	0.5840	0.6016	0.5991
1.1293	16.0	654	0.9826	0.5714	0.5548	0.6000	0.5714
1.0706	16.9786	694	0.9465	0.5945	0.5739	0.6001	0.5945
1.0825	17.9817	735	0.9268	0.6083	0.5861	0.6083	0.6083
1.0989	18.9847	776	0.9349	0.6083	0.5900	0.6168	0.6083
1.0541	19.9878	817	0.9408	0.6175	0.6103	0.6332	0.6175
1.0883	20.9908	858	0.9108	0.6221	0.6116	0.6170	0.6221
1.0828	21.9939	899	0.9412	0.5991	0.5913	0.6205	0.5991
1.0492	22.9969	940	0.9263	0.6129	0.6003	0.6442	0.6129
1.0486	24.0	981	0.9254	0.6129	0.6069	0.6137	0.6129
1.0648	24.9786	1021	0.9165	0.5991	0.5880	0.6002	0.5991
1.079	25.9817	1062	0.9294	0.5899	0.5795	0.6006	0.5899
1.0459	26.9847	1103	0.9237	0.6083	0.5953	0.6168	0.6083
1.057	27.9878	1144	0.9233	0.6083	0.5953	0.6168	0.6083
1.0496	28.9908	1185	0.9237	0.5991	0.5874	0.6077	0.5991
1.0509	29.3578	1200	0.9238	0.5991	0.5874	0.6077	0.5991

Framework versions

• Transformers 4.44.2
• Pytorch 2.4.0+cu121
• Datasets 2.21.0
• Tokenizers 0.19.1