vit-base-food101 / README.md
eslamxm's picture
update model card README.md
f99f99a
metadata
license: apache-2.0
tags:
  - image-classification
  - generated_from_trainer
datasets:
  - food101
metrics:
  - accuracy
model-index:
  - name: vit-base-food101-demo-v5
    results:
      - task:
          name: Image Classification
          type: image-classification
        dataset:
          name: food101
          type: food101
          config: default
          split: train
          args: default
        metrics:
          - name: Accuracy
            type: accuracy
            value: 0.8539405940594059

vit-base-food101-demo-v5

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: 0.5493
  • Accuracy: 0.8539

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
  • mixed_precision_training: Native AMP

Training results

Training Loss Epoch Step Validation Loss Accuracy
1.657 1.0 4735 0.9732 0.7459
0.9869 2.0 9470 0.7987 0.7884
0.71 3.0 14205 0.6364 0.8311
0.4961 4.0 18940 0.5595 0.8487

Framework versions

  • Transformers 4.21.1
  • Pytorch 1.12.0+cu113
  • Datasets 2.4.0
  • Tokenizers 0.12.1