banknote18k / README.md
Jakir057's picture
Model save
0b0c6a5
metadata
license: apache-2.0
base_model: google/vit-base-patch16-224-in21k
tags:
  - generated_from_trainer
metrics:
  - accuracy
model-index:
  - name: banknote18k
    results: []

banknote18k

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: 0.0096
  • Accuracy: 0.9987

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

Training results

Training Loss Epoch Step Validation Loss Accuracy
0.4947 0.12 100 0.3407 0.9451
0.423 0.23 200 0.2200 0.9451
0.2237 0.35 300 0.1613 0.9536
0.2806 0.46 400 0.0884 0.9810
0.1188 0.58 500 0.0512 0.9895
0.3279 0.7 600 0.0568 0.9876
0.1054 0.81 700 0.0342 0.9928
0.0924 0.93 800 0.0536 0.9863
0.1068 1.05 900 0.0746 0.9804
0.213 1.16 1000 0.0340 0.9948
0.159 1.28 1100 0.0426 0.9882
0.1048 1.39 1200 0.0248 0.9948
0.1493 1.51 1300 0.0154 0.9974
0.1274 1.63 1400 0.0394 0.9922
0.0915 1.74 1500 0.0422 0.9882
0.0598 1.86 1600 0.0219 0.9948
0.1241 1.97 1700 0.0173 0.9948
0.1249 2.09 1800 0.0179 0.9954
0.0131 2.21 1900 0.0124 0.9961
0.0392 2.32 2000 0.0123 0.9967
0.0655 2.44 2100 0.0223 0.9948
0.0355 2.56 2200 0.0256 0.9941
0.0335 2.67 2300 0.0147 0.9967
0.0618 2.79 2400 0.0123 0.9974
0.0476 2.9 2500 0.0110 0.9980
0.0452 3.02 2600 0.0192 0.9967
0.0104 3.14 2700 0.0184 0.9967
0.036 3.25 2800 0.0122 0.9974
0.0358 3.37 2900 0.0104 0.9987
0.054 3.48 3000 0.0101 0.9987
0.0395 3.6 3100 0.0132 0.9967
0.0367 3.72 3200 0.0096 0.9987
0.0261 3.83 3300 0.0101 0.9980
0.0017 3.95 3400 0.0096 0.9987

Framework versions

  • Transformers 4.32.1
  • Pytorch 2.0.1+cu118
  • Datasets 2.14.4
  • Tokenizers 0.13.3