jolual2747's picture
VIT model tuned
c4e28b1 verified
|
raw
history blame
2.33 kB
metadata
license: apache-2.0
base_model: google/vit-base-patch16-224-in21k
tags:
  - image-classification
  - clothes-classification
  - generated_from_trainer
metrics:
  - accuracy
model-index:
  - name: vit-clothes-classification
    results: []

vit-clothes-classification

This model is a fine-tuned version of google/vit-base-patch16-224-in21k on the DBQ/Matches.Fashion.Product.prices.France dataset. It achieves the following results on the evaluation set:

  • Loss: 1.2328
  • Accuracy: 0.6395

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

Training results

Training Loss Epoch Step Validation Loss Accuracy
1.0975 0.5714 500 1.2619 0.6111
0.8315 1.1429 1000 1.3133 0.6322
0.7266 1.7143 1500 1.2077 0.6356
0.5451 2.2857 2000 1.2895 0.6556
0.4287 2.8571 2500 1.2736 0.6644
0.2554 3.4286 3000 1.3801 0.6767
0.2265 4.0 3500 1.4924 0.6656
0.0738 4.5714 4000 1.6321 0.68
0.0761 5.1429 4500 1.6676 0.6767
0.0251 5.7143 5000 1.6911 0.7056
0.0147 6.2857 5500 1.7312 0.7
0.0051 6.8571 6000 1.7282 0.6922
0.0028 7.4286 6500 1.7679 0.6967
0.0017 8.0 7000 1.7642 0.6989

Framework versions

  • Transformers 4.40.0
  • Pytorch 2.2.1+cu121
  • Datasets 2.19.0
  • Tokenizers 0.19.1