prototorch_models/examples/knn_iris.py

82 lines
1.9 KiB
Python
Raw Normal View History

2021-06-04 20:21:28 +00:00
"""k-NN example using the Iris dataset from scikit-learn."""
2021-05-11 15:22:02 +00:00
2021-05-21 15:55:55 +00:00
import argparse
import logging
import warnings
2021-05-21 15:55:55 +00:00
import prototorch as pt
2021-05-11 15:22:02 +00:00
import pytorch_lightning as pl
import torch
from prototorch.models import KNN, VisGLVQ2D
from pytorch_lightning.utilities.warnings import PossibleUserWarning
2021-05-21 15:55:55 +00:00
from sklearn.datasets import load_iris
2021-09-10 17:19:51 +00:00
from sklearn.model_selection import train_test_split
from torch.utils.data import DataLoader
warnings.filterwarnings("ignore", category=PossibleUserWarning)
2021-05-11 15:22:02 +00:00
if __name__ == "__main__":
2021-05-21 15:55:55 +00:00
# Command-line arguments
parser = argparse.ArgumentParser()
parser = pl.Trainer.add_argparse_args(parser)
args = parser.parse_args()
2021-05-11 15:22:02 +00:00
# Dataset
2021-09-10 17:19:51 +00:00
X, y = load_iris(return_X_y=True)
X = X[:, 0:3:2]
2021-09-10 17:19:51 +00:00
X_train, X_test, y_train, y_test = train_test_split(
X,
y,
test_size=0.5,
random_state=42,
)
2021-09-10 17:19:51 +00:00
train_ds = pt.datasets.NumpyDataset(X_train, y_train)
test_ds = pt.datasets.NumpyDataset(X_test, y_test)
2021-05-11 15:22:02 +00:00
# Dataloaders
train_loader = DataLoader(train_ds, batch_size=16)
test_loader = DataLoader(test_ds, batch_size=16)
2021-05-11 15:22:02 +00:00
# Hyperparameters
2021-06-04 20:21:28 +00:00
hparams = dict(k=5)
2021-05-11 15:22:02 +00:00
# Initialize the model
model = KNN(hparams, data=train_ds)
2021-05-11 15:22:02 +00:00
2021-06-04 20:21:28 +00:00
# Compute intermediate input and output sizes
model.example_input_array = torch.zeros(4, 2)
# Summary
logging.info(model)
2021-06-04 20:21:28 +00:00
2021-05-11 15:22:02 +00:00
# Callbacks
vis = VisGLVQ2D(
2021-09-10 17:19:51 +00:00
data=(X_train, y_train),
2021-06-04 20:21:28 +00:00
resolution=200,
block=True,
)
2021-05-11 15:22:02 +00:00
# Setup trainer
2021-05-21 15:55:55 +00:00
trainer = pl.Trainer.from_argparse_args(
args,
2021-06-04 20:21:28 +00:00
max_epochs=1,
callbacks=[
vis,
],
log_every_n_steps=1,
detect_anomaly=True,
2021-05-21 15:55:55 +00:00
)
2021-05-11 15:22:02 +00:00
# Training loop
# This is only for visualization. k-NN has no training phase.
trainer.fit(model, train_loader)
# Recall
2021-09-10 17:19:51 +00:00
y_pred = model.predict(torch.tensor(X_train))
logging.info(y_pred)
2021-09-10 17:19:51 +00:00
# Test
trainer.test(model, dataloaders=test_loader)