vit-base-patch16-224-finetuned-algae-wirs

This model is a fine-tuned version of google/vit-base-patch16-224 on the None dataset. It achieves the following results on the evaluation set:

• Loss: 0.9663
• Accuracy: 0.6021

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: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_ratio: 0.1
• num_epochs: 20

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
1.0733	1.0	120	1.0611	0.5781
1.0243	2.0	240	1.0628	0.5663
0.9852	3.0	360	1.0083	0.5845
0.94	4.0	480	1.0005	0.5933
0.9744	5.0	600	1.0102	0.5786
0.9623	6.0	720	0.9840	0.5763
0.9021	7.0	840	0.9869	0.5798
0.9181	8.0	960	0.9755	0.5827
0.8774	9.0	1080	0.9808	0.5798
0.8294	10.0	1200	0.9663	0.6021
0.8015	11.0	1320	0.9739	0.5980
0.8063	12.0	1440	0.9811	0.6009
0.7857	13.0	1560	0.9833	0.5933
0.7085	14.0	1680	0.9887	0.5998
0.7414	15.0	1800	0.9928	0.5974
0.7442	16.0	1920	0.9963	0.5992
0.7142	17.0	2040	1.0041	0.6004
0.7488	18.0	2160	1.0034	0.5962
0.6731	19.0	2280	1.0055	0.6021
0.6905	20.0	2400	1.0033	0.6009

Framework versions

• Transformers 4.26.1
• Pytorch 1.13.1+cu116
• Datasets 2.9.0
• Tokenizers 0.13.2