vit-base-beans

This model is a fine-tuned version of timm/resnet101.a1_in1k on the beans dataset. It achieves the following results on the evaluation set:

• Loss: 0.4913
• Accuracy: 0.8571

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	Validation Loss	Accuracy
1.07	1.0	130	1.0683	0.4135
1.0523	2.0	260	1.0356	0.6241
1.0439	3.0	390	1.0045	0.6617
1.0056	4.0	520	0.9671	0.7293
0.9853	5.0	650	0.9245	0.7895
0.9581	6.0	780	0.8744	0.7820
0.9044	7.0	910	0.8172	0.7820
0.869	8.0	1040	0.7737	0.8271
0.8804	9.0	1170	0.7098	0.8271
0.7757	10.0	1300	0.6705	0.8120
0.7694	11.0	1430	0.6382	0.8571
0.7966	12.0	1560	0.6088	0.7895
0.7425	13.0	1690	0.5724	0.8496
0.7698	14.0	1820	0.5665	0.8195
0.6632	15.0	1950	0.5308	0.8571
0.6162	16.0	2080	0.5262	0.8346
0.6128	17.0	2210	0.5081	0.8421
0.685	18.0	2340	0.4913	0.8571
0.6614	19.0	2470	0.4937	0.8496
0.6934	20.0	2600	0.5027	0.8571

Framework versions

• Transformers 4.47.0.dev0
• Pytorch 2.5.0+cu118
• Datasets 2.21.0
• Tokenizers 0.20.0