vit-base-food101

This model is a fine-tuned version of google/vit-base-patch16-224 on the ethz/food101 dataset. It achieves the following results on the evaluation set:

• Loss: 0.7395
• Accuracy: 0.8017

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: 0.0001
• train_batch_size: 16
• eval_batch_size: 8
• seed: 42
• 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: 8

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
2.5327	0.1320	500	2.3914	0.5946
1.5713	0.2640	1000	1.5558	0.6978
1.2869	0.3960	1500	1.2575	0.7271
1.1479	0.5280	2000	1.1093	0.7476
1.0838	0.6600	2500	1.0286	0.7571
0.9623	0.7920	3000	0.9798	0.7641
0.9855	0.9240	3500	0.9395	0.7670
0.9263	1.0560	4000	0.9113	0.7723
0.8691	1.1880	4500	0.8844	0.7782
0.8025	1.3200	5000	0.8694	0.7768
0.7783	1.4520	5500	0.8574	0.7820
0.7774	1.5839	6000	0.8457	0.7799
0.7716	1.7159	6500	0.8309	0.7871
0.8445	1.8479	7000	0.8230	0.7868
0.8214	1.9799	7500	0.8107	0.7902
0.7226	2.1119	8000	0.8077	0.7897
0.7712	2.2439	8500	0.8015	0.7914
0.7306	2.3759	9000	0.7970	0.7889
0.6829	2.5079	9500	0.7919	0.7912
0.7593	2.6399	10000	0.7883	0.7901
0.6856	2.7719	10500	0.7802	0.7943
0.7156	2.9039	11000	0.7765	0.7976
0.6688	3.0359	11500	0.7735	0.7978
0.6245	3.1679	12000	0.7711	0.7972
0.668	3.2999	12500	0.7679	0.7989
0.6732	3.4319	13000	0.7657	0.7985
0.686	3.5639	13500	0.7645	0.7982
0.7121	3.6959	14000	0.7612	0.7984
0.6513	3.8279	14500	0.7599	0.7993
0.6963	3.9599	15000	0.7585	0.7993
0.7219	4.0919	15500	0.7554	0.7999
0.6253	4.2239	16000	0.7526	0.8016
0.6278	4.3559	16500	0.7504	0.8026
0.6605	4.4879	17000	0.7502	0.8028
0.6447	4.6199	17500	0.7493	0.8028
0.6469	4.7518	18000	0.7463	0.8040
0.6745	4.8838	18500	0.7462	0.8028
0.5882	5.0158	19000	0.7463	0.7995
0.6241	5.1478	19500	0.7428	0.8046
0.62	5.2798	20000	0.7439	0.8013
0.6435	5.4118	20500	0.7422	0.8018
0.6273	5.5438	21000	0.7418	0.8030
0.623	5.6758	21500	0.7415	0.8050
0.6181	5.8078	22000	0.7385	0.8055
0.6382	5.9398	22500	0.7388	0.8071
0.587	6.0718	23000	0.7379	0.8058
0.603	6.2038	23500	0.7374	0.8038
0.6334	6.3358	24000	0.7366	0.8054
0.613	6.4678	24500	0.7364	0.8048
0.5917	6.5998	25000	0.7355	0.8051
0.6167	6.7318	25500	0.7352	0.8059
0.6121	6.8638	26000	0.7347	0.8066
0.6133	6.9958	26500	0.7342	0.8059
0.6304	7.1278	27000	0.7338	0.8057
0.6041	7.2598	27500	0.7342	0.8063
0.6333	7.3918	28000	0.7334	0.8059
0.6234	7.5238	28500	0.7335	0.8061
0.5961	7.6558	29000	0.7334	0.8073
0.61	7.7878	29500	0.7333	0.8070
0.6586	7.9197	30000	0.7331	0.8070

Framework versions

• Transformers 4.50.0
• Pytorch 2.6.0
• Datasets 3.4.1
• Tokenizers 0.21.1