beit-base-patch16-224-finetuned-galaxy10-decals

This model is a fine-tuned version of microsoft/beit-base-patch16-224 on the matthieulel/galaxy10_decals dataset. It achieves the following results on the evaluation set:

• Loss: 0.4611
• Accuracy: 0.8687
• Precision: 0.8668
• Recall: 0.8687
• F1: 0.8672

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: 64
• eval_batch_size: 64
• seed: 42
• gradient_accumulation_steps: 4
• total_train_batch_size: 256
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_ratio: 0.1
• num_epochs: 30

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	Precision	Recall	F1
1.1109	0.99	62	0.8733	0.7108	0.7055	0.7108	0.6891
0.7936	2.0	125	0.6203	0.7903	0.7882	0.7903	0.7816
0.7011	2.99	187	0.5244	0.8269	0.8322	0.8269	0.8246
0.5824	4.0	250	0.5032	0.8298	0.8327	0.8298	0.8264
0.5333	4.99	312	0.5270	0.8236	0.8299	0.8236	0.8246
0.5432	6.0	375	0.5006	0.8337	0.8418	0.8337	0.8286
0.5206	6.99	437	0.4891	0.8331	0.8357	0.8331	0.8330
0.4597	8.0	500	0.4484	0.8512	0.8510	0.8512	0.8501
0.4779	8.99	562	0.4499	0.8529	0.8547	0.8529	0.8518
0.4199	10.0	625	0.4354	0.8602	0.8583	0.8602	0.8576
0.4242	10.99	687	0.4295	0.8512	0.8523	0.8512	0.8510
0.4048	12.0	750	0.4662	0.8461	0.8490	0.8461	0.8456
0.403	12.99	812	0.4424	0.8591	0.8586	0.8591	0.8579
0.3958	14.0	875	0.4448	0.8546	0.8558	0.8546	0.8546
0.3602	14.99	937	0.4378	0.8585	0.8582	0.8585	0.8572
0.352	16.0	1000	0.4570	0.8546	0.8539	0.8546	0.8527
0.3041	16.99	1062	0.4601	0.8506	0.8486	0.8506	0.8486
0.3439	18.0	1125	0.4483	0.8489	0.8519	0.8489	0.8496
0.3148	18.99	1187	0.4372	0.8625	0.8598	0.8625	0.8606
0.3075	20.0	1250	0.4370	0.8585	0.8556	0.8585	0.8567
0.2999	20.99	1312	0.4570	0.8546	0.8540	0.8546	0.8535
0.3156	22.0	1375	0.4541	0.8540	0.8525	0.8540	0.8527
0.2845	22.99	1437	0.4702	0.8574	0.8584	0.8574	0.8576
0.2702	24.0	1500	0.4624	0.8630	0.8621	0.8630	0.8615
0.2914	24.99	1562	0.4670	0.8591	0.8575	0.8591	0.8577
0.2464	26.0	1625	0.4581	0.8658	0.8648	0.8658	0.8650
0.2621	26.99	1687	0.4688	0.8658	0.8662	0.8658	0.8657
0.2731	28.0	1750	0.4664	0.8664	0.8643	0.8664	0.8648
0.2448	28.99	1812	0.4611	0.8687	0.8668	0.8687	0.8672
0.2517	29.76	1860	0.4612	0.8675	0.8660	0.8675	0.8663

Framework versions

• Transformers 4.37.2
• Pytorch 2.3.0
• Datasets 2.19.1
• Tokenizers 0.15.1