food_classifier_2025_03_18_20_39

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

• Loss: 0.5122
• Accuracy: 0.8746

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: 0.0006
• train_batch_size: 128
• eval_batch_size: 128
• seed: 42
• distributed_type: multi-GPU
• num_devices: 4
• gradient_accumulation_steps: 4
• total_train_batch_size: 2048
• total_eval_batch_size: 512
• optimizer: Use OptimizerNames.ADAMW_TORCH with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: linear
• lr_scheduler_warmup_ratio: 0.1
• num_epochs: 18

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
3.9399	1.0	37	3.1445	0.7088
1.578	2.0	74	1.1087	0.7971
0.9126	3.0	111	0.7459	0.8190
0.7204	4.0	148	0.6649	0.8352
0.611	5.0	185	0.6167	0.8424
0.5583	6.0	222	0.5946	0.8468
0.4702	7.0	259	0.5649	0.8561
0.4427	8.0	296	0.5751	0.8512
0.3757	9.0	333	0.5720	0.8535
0.3356	10.0	370	0.5514	0.8589
0.3129	11.0	407	0.5458	0.8612
0.2894	12.0	444	0.5399	0.8595
0.2513	13.0	481	0.5293	0.8675
0.2419	14.0	518	0.5299	0.868
0.2137	15.0	555	0.5250	0.8703
0.2215	16.0	592	0.5194	0.8676
0.2046	17.0	629	0.5201	0.8689
0.1864	18.0	666	0.5122	0.8746

Framework versions

• Transformers 4.49.0
• Pytorch 2.6.0+cu124
• Datasets 3.3.2
• Tokenizers 0.21.0