Jakir057/finetuned-indian-food - AI Model Zoo

finetuned-indian-food

This model is a fine-tuned version of google/vit-base-patch16-224-in21k on the indian_food_images dataset. It achieves the following results on the evaluation set:

Loss: 0.0026
Accuracy: 0.9996

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: 0.0002
train_batch_size: 16
eval_batch_size: 8
seed: 42
optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
lr_scheduler_type: linear
num_epochs: 4

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
0.7056	0.1	100	0.5113	0.8881
0.3027	0.21	200	0.1280	0.9796
0.2823	0.31	300	0.1580	0.9656
0.3273	0.42	400	0.0879	0.9837
0.1808	0.52	500	0.0812	0.9822
0.2101	0.63	600	0.0339	0.9937
0.1495	0.73	700	0.0568	0.9833
0.1296	0.84	800	0.0629	0.9844
0.1462	0.94	900	0.0886	0.9733
0.0519	1.04	1000	0.0544	0.9870
0.3192	1.15	1100	0.0892	0.9726
0.158	1.25	1200	0.0632	0.98
0.0266	1.36	1300	0.0233	0.9944
0.1832	1.46	1400	0.0292	0.9930
0.1212	1.57	1500	0.0489	0.9852
0.0994	1.67	1600	0.0142	0.9974
0.0219	1.78	1700	0.0277	0.9930
0.0664	1.88	1800	0.0158	0.9974
0.0834	1.99	1900	0.0124	0.9978
0.1093	2.09	2000	0.0140	0.9974
0.1726	2.19	2100	0.0147	0.9963
0.0476	2.3	2200	0.0058	0.9993
0.0257	2.4	2300	0.0424	0.9911
0.0215	2.51	2400	0.0076	0.9989
0.0748	2.61	2500	0.0099	0.9974
0.0059	2.72	2600	0.0053	0.9993
0.0527	2.82	2700	0.0149	0.9963
0.0203	2.93	2800	0.0041	0.9993
0.0791	3.03	2900	0.0033	0.9989
0.0389	3.13	3000	0.0033	0.9989
0.0459	3.24	3100	0.0044	0.9989
0.0276	3.34	3200	0.0031	0.9996
0.0139	3.45	3300	0.0028	0.9996
0.0076	3.55	3400	0.0055	0.9985
0.0097	3.66	3500	0.0027	0.9996
0.0193	3.76	3600	0.0026	0.9996
0.0471	3.87	3700	0.0027	0.9996
0.0282	3.97	3800	0.0027	0.9996

Framework versions

Transformers 4.32.1
Pytorch 2.0.1+cu118
Datasets 2.14.4
Tokenizers 0.13.3