prototorch_models/examples/glvq_iris.py

134 lines
4.3 KiB
Python
Raw Normal View History

2021-04-21 12:54:07 +00:00
"""GLVQ example using the Iris dataset."""
import argparse
2021-04-21 12:54:07 +00:00
import numpy as np
import pytorch_lightning as pl
import torch
from matplotlib import pyplot as plt
from prototorch.models.glvq import GLVQ
from sklearn.datasets import load_iris
from torch.utils.data import DataLoader, TensorDataset
class NumpyDataset(TensorDataset):
def __init__(self, *arrays):
tensors = [torch.from_numpy(arr) for arr in arrays]
super().__init__(*tensors)
2021-04-21 19:35:52 +00:00
class GLVQIris(GLVQ):
@staticmethod
def add_model_specific_args(parent_parser):
parser = argparse.ArgumentParser(parents=[parent_parser],
add_help=False)
parser.add_argument("--epochs", type=int, default=1)
parser.add_argument("--lr", type=float, default=1e-1)
parser.add_argument("--batch_size", type=int, default=150)
parser.add_argument("--input_dim", type=int, default=2)
parser.add_argument("--nclasses", type=int, default=3)
2021-04-21 19:35:52 +00:00
parser.add_argument("--prototypes_per_class", type=int, default=3)
parser.add_argument("--prototype_initializer",
type=str,
default="stratified_mean")
return parser
2021-04-21 12:54:07 +00:00
class VisualizationCallback(pl.Callback):
def __init__(self,
x_train,
y_train,
title="Prototype Visualization",
cmap="viridis"):
super().__init__()
self.x_train = x_train
self.y_train = y_train
self.title = title
self.fig = plt.figure(self.title)
self.cmap = cmap
def on_epoch_end(self, trainer, pl_module):
protos = pl_module.prototypes
plabels = pl_module.prototype_labels
ax = self.fig.gca()
ax.cla()
ax.set_title(self.title)
ax.set_xlabel("Data dimension 1")
ax.set_ylabel("Data dimension 2")
ax.scatter(x_train[:, 0], x_train[:, 1], c=y_train, edgecolor="k")
ax.scatter(protos[:, 0],
protos[:, 1],
c=plabels,
cmap=self.cmap,
edgecolor="k",
marker="D",
s=50)
x = np.vstack((x_train, protos))
x_min, x_max = x[:, 0].min() - 1, x[:, 0].max() + 1
y_min, y_max = x[:, 1].min() - 1, x[:, 1].max() + 1
xx, yy = np.meshgrid(np.arange(x_min, x_max, 1 / 50),
np.arange(y_min, y_max, 1 / 50))
mesh_input = np.c_[xx.ravel(), yy.ravel()]
y_pred = pl_module.predict(torch.Tensor(mesh_input))
y_pred = y_pred.reshape(xx.shape)
ax.contourf(xx, yy, y_pred, cmap=self.cmap, alpha=0.35)
ax.set_xlim(left=x_min + 0, right=x_max - 0)
ax.set_ylim(bottom=y_min + 0, top=y_max - 0)
plt.pause(0.1)
if __name__ == "__main__":
2021-04-21 19:35:52 +00:00
# For best-practices when using `argparse` with `pytorch_lightning`, see
# https://pytorch-lightning.readthedocs.io/en/stable/common/hyperparameters.html
2021-04-21 19:35:52 +00:00
parser = argparse.ArgumentParser()
2021-04-21 13:52:42 +00:00
# Dataset
2021-04-21 12:54:07 +00:00
x_train, y_train = load_iris(return_X_y=True)
x_train = x_train[:, [0, 2]]
train_ds = NumpyDataset(x_train, y_train)
2021-04-21 13:52:42 +00:00
# Dataloaders
2021-04-21 12:54:07 +00:00
train_loader = DataLoader(train_ds, num_workers=0, batch_size=150)
2021-04-21 19:35:52 +00:00
# Add model specific args
parser = GLVQIris.add_model_specific_args(parser)
2021-04-21 12:54:07 +00:00
2021-04-21 13:52:42 +00:00
# Callbacks
2021-04-21 12:54:07 +00:00
vis = VisualizationCallback(x_train, y_train)
2021-04-21 13:52:42 +00:00
2021-04-21 19:35:52 +00:00
# Automatically add trainer-specific-args like `--gpus`, `--num_nodes` etc.
parser = pl.Trainer.add_argparse_args(parser)
2021-04-21 13:52:42 +00:00
# Setup trainer
2021-04-21 19:35:52 +00:00
trainer = pl.Trainer.from_argparse_args(
parser,
max_epochs=10,
callbacks=[
vis, # comment this line out to disable the visualization
],
)
2021-04-21 19:35:52 +00:00
# trainer.tune(model)
# Initialize the model
args = parser.parse_args()
model = GLVQIris(args, data=[x_train, y_train])
2021-04-21 19:35:52 +00:00
# Model summary
print(model)
2021-04-21 12:54:07 +00:00
2021-04-21 13:52:42 +00:00
# Training loop
2021-04-21 12:54:07 +00:00
trainer.fit(model, train_loader)
# Save the model manually (use `pl.callbacks.ModelCheckpoint` to automate)
ckpt = "glvq_iris.ckpt"
trainer.save_checkpoint(ckpt)
# Load the checkpoint
2021-04-21 19:35:52 +00:00
new_model = GLVQIris.load_from_checkpoint(checkpoint_path=ckpt)
print(new_model)
# Continue training
trainer.fit(new_model, train_loader) # TODO See why this fails!