levit-256-finetuned-flower

This model is a fine-tuned version of facebook/levit-256 on the imagefolder dataset. It achieves the following results on the evaluation set:

• Loss: 0.1677
• Accuracy: 0.9521
• Precision: 0.9523
• Recall: 0.9521
• F1: 0.9518

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
0.599	1.0	40	0.5907	0.8207	0.8515	0.8207	0.8219
0.7842	2.0	80	1.4800	0.6693	0.7271	0.6693	0.6607
0.7716	3.0	120	0.8614	0.7554	0.7853	0.7554	0.7544
0.5976	4.0	160	0.5576	0.8243	0.8470	0.8243	0.8260
0.488	5.0	200	0.4656	0.8555	0.8724	0.8555	0.8546
0.4871	6.0	240	0.4387	0.8672	0.8823	0.8672	0.8672
0.3606	7.0	280	0.3041	0.9045	0.9053	0.9045	0.9034
0.3159	8.0	320	0.3283	0.8976	0.9022	0.8976	0.8961
0.3078	9.0	360	0.2848	0.9125	0.9156	0.9125	0.9124
0.2922	10.0	400	0.2526	0.9180	0.9212	0.9180	0.9184
0.2412	11.0	440	0.2367	0.9281	0.9306	0.9281	0.9280
0.2095	12.0	480	0.2283	0.9314	0.9323	0.9314	0.9305
0.1786	13.0	520	0.1890	0.9408	0.9412	0.9408	0.9408
0.123	14.0	560	0.2071	0.9383	0.9398	0.9383	0.9382
0.1481	15.0	600	0.1854	0.9426	0.9433	0.9426	0.9426
0.125	16.0	640	0.2051	0.9376	0.9400	0.9376	0.9373
0.1135	17.0	680	0.1785	0.9495	0.9496	0.9495	0.9495
0.0815	18.0	720	0.1655	0.9539	0.9542	0.9539	0.9538
0.0784	19.0	760	0.1707	0.9525	0.9527	0.9525	0.9521
0.0905	20.0	800	0.1677	0.9521	0.9523	0.9521	0.9518

Framework versions

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