finetuned-fake-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: 0.4855
• Accuracy: 0.8548

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: 3e-05
• train_batch_size: 4
• eval_batch_size: 8
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: cosine
• num_epochs: 15
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
0.6061	1.0	176	0.5937	0.6855
0.481	2.0	352	0.5138	0.8226
0.5522	3.0	528	0.4973	0.8065
0.4092	4.0	704	0.5557	0.7903
0.4882	5.0	880	0.4998	0.7984
0.4442	6.0	1056	0.4647	0.8387
0.5749	7.0	1232	0.4464	0.8306
0.4529	8.0	1408	0.5366	0.8065
0.5287	9.0	1584	0.4633	0.8387
0.3821	10.0	1760	0.4983	0.8387
0.2409	11.0	1936	0.4855	0.8548
0.2025	12.0	2112	0.5102	0.8387
0.2045	13.0	2288	0.4942	0.8387
0.4097	14.0	2464	0.4954	0.8387
0.5798	15.0	2640	0.4941	0.8387

Framework versions

• Transformers 4.44.2
• Pytorch 2.4.1+cu121
• Datasets 3.0.1
• Tokenizers 0.19.1