cvt-13-finetuned-flower

This model is a fine-tuned version of microsoft/cvt-13 on the imagefolder dataset. It achieves the following results on the evaluation set:

• Loss: 0.2151
• Accuracy: 0.9368
• Precision: 0.9375
• Recall: 0.9368
• F1: 0.9342

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: 0.005
• 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: 20

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	Precision	Recall	F1
1.0555	1.0	40	0.3933	0.8766	0.8828	0.8766	0.8713
1.1941	2.0	80	1.0797	0.6726	0.7515	0.6726	0.6546
1.2286	3.0	120	0.8459	0.7347	0.7820	0.7347	0.7343
1.209	4.0	160	0.6660	0.7880	0.8173	0.7880	0.7833
1.1158	5.0	200	0.7348	0.7597	0.7809	0.7597	0.7561
1.1113	6.0	240	0.6387	0.8062	0.8164	0.8062	0.7986
1.0332	7.0	280	0.6555	0.7887	0.8064	0.7887	0.7831
1.0234	8.0	320	0.5776	0.8276	0.8447	0.8276	0.8177
0.9997	9.0	360	0.5784	0.8214	0.8421	0.8214	0.8169
0.9421	10.0	400	0.4667	0.8486	0.8600	0.8486	0.8453
0.9057	11.0	440	0.4508	0.8541	0.8711	0.8541	0.8487
0.8662	12.0	480	0.3517	0.8911	0.8938	0.8911	0.8868
0.8341	13.0	520	0.3191	0.8976	0.9021	0.8976	0.8945
0.757	14.0	560	0.2785	0.9183	0.9199	0.9183	0.9144
0.7906	15.0	600	0.2698	0.9201	0.9218	0.9201	0.9172
0.7464	16.0	640	0.2594	0.9216	0.9232	0.9216	0.9188
0.7335	17.0	680	0.2491	0.9263	0.9281	0.9263	0.9240
0.7085	18.0	720	0.2396	0.9303	0.9304	0.9303	0.9272
0.7177	19.0	760	0.2171	0.9350	0.9355	0.9350	0.9321
0.6735	20.0	800	0.2151	0.9368	0.9375	0.9368	0.9342

Framework versions

• Transformers 4.39.3
• Pytorch 2.0.1
• Datasets 2.18.0
• Tokenizers 0.15.2