koobear/wbc-100-pretrain-20-epoch - AI Model Zoo

wbc-100-pretrain-20-epoch

This model is a fine-tuned version of koobear/masked-pretraining-20-epoch on the image_folder dataset. It achieves the following results on the evaluation set:

Loss: 0.0493
Accuracy: 0.9884
F1: 0.9884
Precision: 0.9884
Recall: 0.9884
Balanced Acc: 0.9762

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: 2e-05
train_batch_size: 64
eval_batch_size: 64
seed: 42
optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
lr_scheduler_type: linear
num_epochs: 20

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	F1	Precision	Recall	Balanced Acc
0.5865	1.0	132	0.2601	0.9606	0.9609	0.9634	0.9606	0.9431
0.2703	2.0	264	0.1383	0.9821	0.9821	0.9822	0.9821	0.9750
0.1952	3.0	396	0.1149	0.9809	0.9812	0.9818	0.9809	0.9811
0.18	4.0	528	0.0914	0.9873	0.9872	0.9873	0.9873	0.9771
0.1475	5.0	660	0.0819	0.9873	0.9874	0.9878	0.9873	0.9837
0.1556	6.0	792	0.0796	0.9855	0.9857	0.9861	0.9855	0.9830
0.1276	7.0	924	0.0746	0.9878	0.9878	0.9879	0.9878	0.9728
0.1205	8.0	1056	0.0689	0.9873	0.9872	0.9874	0.9873	0.9702
0.1196	9.0	1188	0.0680	0.9855	0.9856	0.9862	0.9855	0.9775
0.1111	10.0	1320	0.0605	0.9873	0.9873	0.9875	0.9873	0.9832
0.1052	11.0	1452	0.0522	0.9902	0.9902	0.9902	0.9902	0.9845
0.1015	12.0	1584	0.0545	0.9896	0.9896	0.9896	0.9896	0.9809
0.1044	13.0	1716	0.0529	0.9896	0.9896	0.9896	0.9896	0.9780
0.097	14.0	1848	0.0521	0.9878	0.9878	0.9879	0.9878	0.9708
0.0943	15.0	1980	0.0468	0.9902	0.9902	0.9905	0.9902	0.9875
0.0884	16.0	2112	0.0454	0.9919	0.9919	0.9920	0.9919	0.9889
0.0767	17.0	2244	0.0465	0.9896	0.9896	0.9896	0.9896	0.9816
0.0834	18.0	2376	0.0512	0.9896	0.9895	0.9896	0.9896	0.9731
0.0819	19.0	2508	0.0497	0.9890	0.9890	0.9890	0.9890	0.9767
0.0822	20.0	2640	0.0493	0.9884	0.9884	0.9884	0.9884	0.9762

Framework versions

Transformers 4.33.0
Pytorch 2.0.0
Datasets 2.1.0
Tokenizers 0.13.3