vit-base-beans

This model is a fine-tuned version of timm/resnet18.a1_in1k on the beans dataset. It achieves the following results on the evaluation set:

• Loss: 0.6875
• Accuracy: 0.8647

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: 8
• eval_batch_size: 8
• seed: 1337
• optimizer: Use adamw_torch with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: linear
• num_epochs: 20.0

Training results

Training Loss	Epoch	Step	Accuracy	Validation Loss
1.0864	1.0	130	0.4286	1.0878
1.0629	2.0	260	0.5489	1.0594
1.0434	3.0	390	0.6767	1.0230
1.0214	4.0	520	0.6767	0.9965
1.0026	5.0	650	0.7444	0.9569
0.9753	6.0	780	0.7820	0.9288
0.9252	7.0	910	0.7970	0.8875
0.9192	8.0	1040	0.8120	0.8506
0.9008	9.0	1170	0.8045	0.8338
0.8079	10.0	1300	0.8421	0.8104
0.8332	11.0	1430	0.8346	0.7806
0.8103	12.0	1560	0.8346	0.7586
0.8149	13.0	1690	0.8421	0.7571
0.8186	14.0	1820	0.8271	0.7540
0.7929	15.0	1950	0.8120	0.7412
0.774	16.0	2080	0.7370	0.8496
0.7613	17.0	2210	0.7059	0.8496
0.7778	18.0	2340	0.6930	0.8271
0.8081	19.0	2470	0.6890	0.8647
0.7916	20.0	2600	0.6875	0.8647

Framework versions

• Transformers 4.47.0.dev0
• Pytorch 2.4.1+cu118
• Datasets 2.21.0
• Tokenizers 0.20.0