---
license: afl-3.0
tags:
- image-classification
library_name: keras
datasets:
- mnist
metrics:
- accuracy
model-index:
- name: resnet_mnist_digits
  results:
  - task:
      type: image-classification
      name: Image Classification
    dataset:
      type: mnist
      name: MNIST
    metrics:
      - type: accuracy
        value: .9904
        name: Accuracy
        verified: false
---

# Model Card for resnet_mnist_digits

This model is is a Residual Neural Network (ResNet) for classifying handwritten digits in the MNIST dataset.
This model has 1.35 M parameters and achieves 99.04% accuracy on the MNIST test dataset (i.e., on digits not seen during training).

## Model Details

### Model Description

This model takes as an input a 224x224 array of MNIST digits with values normalized to [0, 1].
Intended to compare to 224x224 vision transformers.
The model was trained using Keras on an Nvidia Ampere A100.

- **Developed by:** Phillip Allen Lane
- **Model type:** ResNet
- **License:** afl-3.0

### How to Get Started with the Model

Use the code below to get started with the model.

```py
from tensorflow.keras import models
from tensorflow.keras.datasets import mnist
from tensorflow.keras.utils import to_categorical
from keras.utils.data_utils import get_file

# load the MNIST dataset test images and labels
(_, _), (test_images, test_labels) = mnist.load_data()

# normalize the images
test_images = test_images.astype('float32') / 255
test_images = np.expand_dims(test_images, axis=-1)
test_images = np.repeat(vis_test_images, 3, axis=-1)
test_images = tf.image.resize(vis_test_images, [224,224]).numpy()
# create one-hot labels
test_labels_onehot = to_categorical(test_labels)

# download the model
model_path = get_file('/path/to/large_resnet_mnist.hdf5', 'https://huggingface.co/lane99/resnet_mnist_digits_highres/resolve/main/large-resnet-mnist.hdf5')
# import the model
resnet = models.load_model(model_path)

# evaluate the model
evaluation_conv = resnet.evaluate(test_images[...,0], test_labels_onehot)
print("Accuracy: ", str(evaluation_conv[1]))
```

## Training Details

### Training Data

This model was trained on the 60,000 entries in the MNIST training dataset.

### Training Procedure 

This model was trained with a 0.1 validation split for 10 epochs using a batch size of 128.