prototorch_models/examples/siamese_glvq_iris.py

68 lines
1.9 KiB
Python
Raw Normal View History

2021-04-27 12:35:17 +00:00
"""Siamese GLVQ example using all four dimensions of the Iris dataset."""
import pytorch_lightning as pl
import torch
from prototorch.components import initializers as cinit
2021-04-27 13:38:57 +00:00
from prototorch.datasets.abstract import NumpyDataset
from prototorch.models.callbacks.visualization import VisSiameseGLVQ2D
from prototorch.models.glvq import SiameseGLVQ
from sklearn.datasets import load_iris
from torch.utils.data import DataLoader
2021-04-27 12:35:17 +00:00
class Backbone(torch.nn.Module):
"""Two fully connected layers with ReLU activation."""
2021-04-27 12:35:17 +00:00
def __init__(self, input_size=4, hidden_size=10, latent_size=2):
super().__init__()
self.input_size = input_size
self.hidden_size = hidden_size
self.latent_size = latent_size
self.dense1 = torch.nn.Linear(self.input_size, self.hidden_size)
self.dense2 = torch.nn.Linear(self.hidden_size, self.latent_size)
self.relu = torch.nn.ReLU()
def forward(self, x):
x = self.relu(self.dense1(x))
out = self.relu(self.dense2(x))
return out
2021-04-27 12:35:17 +00:00
if __name__ == "__main__":
# Dataset
x_train, y_train = load_iris(return_X_y=True)
train_ds = NumpyDataset(x_train, y_train)
# Reproducibility
pl.utilities.seed.seed_everything(seed=2)
2021-04-27 12:35:17 +00:00
# Dataloaders
train_loader = DataLoader(train_ds, num_workers=0, batch_size=150)
# Hyperparameters
hparams = dict(
nclasses=3,
prototypes_per_class=2,
prototype_initializer=cinit.SMI(torch.Tensor(x_train),
torch.Tensor(y_train)),
proto_lr=0.001,
bb_lr=0.001,
2021-04-27 12:35:17 +00:00
)
# Initialize the model
model = SiameseGLVQ(
hparams,
backbone_module=Backbone,
)
2021-04-27 12:35:17 +00:00
# Model summary
print(model)
# Callbacks
vis = VisSiameseGLVQ2D(x_train, y_train, border=0.1)
2021-04-27 12:35:17 +00:00
# Setup trainer
trainer = pl.Trainer(max_epochs=100, callbacks=[vis])
# Training loop
trainer.fit(model, train_loader)