fashion-mnist example (#180)

* fashion mnist example

* fix from review
This commit is contained in:
Kashif Rasul 2023-12-23 16:34:45 +01:00 committed by GitHub
parent 848f118ac5
commit 0371d90ccb
No known key found for this signature in database
GPG Key ID: 4AEE18F83AFDEB23
5 changed files with 46 additions and 11 deletions

View File

@ -14,10 +14,16 @@ Run the example with:
python main.py python main.py
``` ```
By default the example runs on the CPU. To run on the GPU, use: By default, the example runs on the CPU. To run on the GPU, use:
``` ```
python main.py --gpu python main.py --gpu
``` ```
For a full list of options run:
```
python main.py --help
```
To run the PyTorch or Jax examples install the respective framework. To run the PyTorch or Jax examples install the respective framework.

View File

@ -54,10 +54,10 @@ if __name__ == "__main__":
batch_size = 256 batch_size = 256
num_epochs = 10 num_epochs = 10
learning_rate = 1e-1 learning_rate = 1e-1
dataset = "mnist"
# Load the data # Load the data
train_images, train_labels, test_images, test_labels = mnist.mnist() train_images, train_labels, test_images, test_labels = getattr(mnist, dataset)()
# Load the model # Load the model
key, subkey = jax.random.split(jax.random.PRNGKey(seed)) key, subkey = jax.random.split(jax.random.PRNGKey(seed))
params = init_model( params = init_model(

View File

@ -45,7 +45,7 @@ def batch_iterate(batch_size, X, y):
yield X[ids], y[ids] yield X[ids], y[ids]
def main(): def main(args):
seed = 0 seed = 0
num_layers = 2 num_layers = 2
hidden_dim = 32 hidden_dim = 32
@ -57,7 +57,9 @@ def main():
np.random.seed(seed) np.random.seed(seed)
# Load the data # Load the data
train_images, train_labels, test_images, test_labels = map(mx.array, mnist.mnist()) train_images, train_labels, test_images, test_labels = map(
mx.array, getattr(mnist, args.dataset)()
)
# Load the model # Load the model
model = MLP(num_layers, train_images.shape[-1], hidden_dim, num_classes) model = MLP(num_layers, train_images.shape[-1], hidden_dim, num_classes)
@ -83,7 +85,14 @@ def main():
if __name__ == "__main__": if __name__ == "__main__":
parser = argparse.ArgumentParser("Train a simple MLP on MNIST with MLX.") parser = argparse.ArgumentParser("Train a simple MLP on MNIST with MLX.")
parser.add_argument("--gpu", action="store_true", help="Use the Metal back-end.") parser.add_argument("--gpu", action="store_true", help="Use the Metal back-end.")
parser.add_argument(
"--dataset",
type=str,
default="mnist",
choices=["mnist", "fashion_mnist"],
help="The dataset to use.",
)
args = parser.parse_args() args = parser.parse_args()
if not args.gpu: if not args.gpu:
mx.set_default_device(mx.cpu) mx.set_default_device(mx.cpu)
main() main(args)

View File

@ -8,7 +8,9 @@ from urllib import request
import numpy as np import numpy as np
def mnist(save_dir="/tmp"): def mnist(
save_dir="/tmp", base_url="http://yann.lecun.com/exdb/mnist/", filename="mnist.pkl"
):
""" """
Load the MNIST dataset in 4 tensors: train images, train labels, Load the MNIST dataset in 4 tensors: train images, train labels,
test images, and test labels. test images, and test labels.
@ -20,7 +22,6 @@ def mnist(save_dir="/tmp"):
""" """
def download_and_save(save_file): def download_and_save(save_file):
base_url = "http://yann.lecun.com/exdb/mnist/"
filename = [ filename = [
["training_images", "train-images-idx3-ubyte.gz"], ["training_images", "train-images-idx3-ubyte.gz"],
["test_images", "t10k-images-idx3-ubyte.gz"], ["test_images", "t10k-images-idx3-ubyte.gz"],
@ -45,13 +46,15 @@ def mnist(save_dir="/tmp"):
with open(save_file, "wb") as f: with open(save_file, "wb") as f:
pickle.dump(mnist, f) pickle.dump(mnist, f)
save_file = os.path.join(save_dir, "mnist.pkl") save_file = os.path.join(save_dir, filename)
if not os.path.exists(save_file): if not os.path.exists(save_file):
download_and_save(save_file) download_and_save(save_file)
with open(save_file, "rb") as f: with open(save_file, "rb") as f:
mnist = pickle.load(f) mnist = pickle.load(f)
preproc = lambda x: x.astype(np.float32) / 255.0 def preproc(x):
return x.astype(np.float32) / 255.0
mnist["training_images"] = preproc(mnist["training_images"]) mnist["training_images"] = preproc(mnist["training_images"])
mnist["test_images"] = preproc(mnist["test_images"]) mnist["test_images"] = preproc(mnist["test_images"])
return ( return (
@ -62,6 +65,14 @@ def mnist(save_dir="/tmp"):
) )
def fashion_mnist(save_dir="/tmp"):
return mnist(
save_dir,
base_url="http://fashion-mnist.s3-website.eu-central-1.amazonaws.com/",
filename="fashion_mnist.pkl",
)
if __name__ == "__main__": if __name__ == "__main__":
train_x, train_y, test_x, test_y = mnist() train_x, train_y, test_x, test_y = mnist()
assert train_x.shape == (60000, 28 * 28), "Wrong training set size" assert train_x.shape == (60000, 28 * 28), "Wrong training set size"

View File

@ -49,6 +49,13 @@ def batch_iterate(batch_size, X, y, device):
if __name__ == "__main__": if __name__ == "__main__":
parser = argparse.ArgumentParser("Train a simple MLP on MNIST with PyTorch.") parser = argparse.ArgumentParser("Train a simple MLP on MNIST with PyTorch.")
parser.add_argument("--gpu", action="store_true", help="Use the Metal back-end.") parser.add_argument("--gpu", action="store_true", help="Use the Metal back-end.")
parser.add_argument(
"--dataset",
type=str,
default="mnist",
choices=["mnist", "fashion_mnist"],
help="The dataset to use.",
)
args = parser.parse_args() args = parser.parse_args()
if not args.gpu: if not args.gpu:
@ -71,7 +78,9 @@ if __name__ == "__main__":
else: else:
return torch.from_numpy(x.astype(int)).to(device) return torch.from_numpy(x.astype(int)).to(device)
train_images, train_labels, test_images, test_labels = map(to_tensor, mnist.mnist()) train_images, train_labels, test_images, test_labels = map(
to_tensor, getattr(mnist, args.dataset)()
)
# Load the model # Load the model
model = MLP(num_layers, train_images.shape[-1], hidden_dim, num_classes).to(device) model = MLP(num_layers, train_images.shape[-1], hidden_dim, num_classes).to(device)