vit-base-patch16-224-in21k-finetuned-lora-food101

This model is a fine-tuned version of google/vit-base-patch16-224-in21k on the food101 dataset. It achieves the following results on the evaluation set:

• Loss: 0.1158
• Accuracy: 0.962

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: 128
• eval_batch_size: 128
• seed: 42
• gradient_accumulation_steps: 4
• total_train_batch_size: 512
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 5

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
No log	1.0	9	0.1536	0.95
0.2548	2.0	18	0.1263	0.96
0.1404	3.0	27	0.1158	0.962
0.1214	4.0	36	0.1232	0.96
0.1238	5.0	45	0.1177	0.96

Framework versions

• PEFT 0.4.0
• Transformers 4.33.2
• Pytorch 2.0.1+cu118
• Datasets 2.14.5
• Tokenizers 0.13.3