julius/prototorch_models
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

Make siamese example script reproducible

Browse Source
This commit is contained in:
Jensun Ravichandran 2021-05-07 13:07:30 +02:00
parent 1b9bcf21f6
commit e87663d10c
1 changed files with 16 additions and 12 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

28
examples/siamese_glvq_iris.py
View File

@ -2,18 +2,16 @@
import pytorch_lightning as pl import pytorch_lightning as pl
import torch import torch
from prototorch.components import ( from prototorch.components import initializers as cinit
StratifiedMeanInitializer
)
from prototorch.datasets.abstract import NumpyDataset 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 sklearn.datasets import load_iris
from torch.utils.data import DataLoader from torch.utils.data import DataLoader
from prototorch.models.callbacks.visualization import VisSiameseGLVQ2D
from prototorch.models.glvq import SiameseGLVQ
class Backbone(torch.nn.Module): class Backbone(torch.nn.Module):
"""Two fully connected layers with ReLU activation."""
def __init__(self, input_size=4, hidden_size=10, latent_size=2): def __init__(self, input_size=4, hidden_size=10, latent_size=2):
super().__init__() super().__init__()
self.input_size = input_size self.input_size = input_size
@ -24,7 +22,9 @@ class Backbone(torch.nn.Module):
self.relu = torch.nn.ReLU() self.relu = torch.nn.ReLU()
def forward(self, x): def forward(self, x):
return self.relu(self.dense2(self.relu(self.dense1(x)))) x = self.relu(self.dense1(x))
out = self.relu(self.dense2(x))
return out
if __name__ == "__main__": if __name__ == "__main__":
@ -32,16 +32,20 @@ if __name__ == "__main__":
x_train, y_train = load_iris(return_X_y=True) x_train, y_train = load_iris(return_X_y=True)
train_ds = NumpyDataset(x_train, y_train) train_ds = NumpyDataset(x_train, y_train)
# Reproducibility
pl.utilities.seed.seed_everything(seed=2)
# Dataloaders # Dataloaders
train_loader = DataLoader(train_ds, num_workers=0, batch_size=150) train_loader = DataLoader(train_ds, num_workers=0, batch_size=150)
# Hyperparameters # Hyperparameters
hparams = dict( hparams = dict(
nclasses=3, nclasses=3,
prototypes_per_class=1, prototypes_per_class=2,
prototype_initializer=StratifiedMeanInitializer( prototype_initializer=cinit.SMI(torch.Tensor(x_train),
torch.Tensor(x_train), torch.Tensor(y_train)), torch.Tensor(y_train)),
lr=0.01, proto_lr=0.001,
bb_lr=0.001,
) )
# Initialize the model # Initialize the model
@ -54,7 +58,7 @@ if __name__ == "__main__":
print(model) print(model)
# Callbacks # Callbacks
vis = VisSiameseGLVQ2D(x_train, y_train) vis = VisSiameseGLVQ2D(x_train, y_train, border=0.1)
# Setup trainer # Setup trainer
trainer = pl.Trainer(max_epochs=100, callbacks=[vis]) trainer = pl.Trainer(max_epochs=100, callbacks=[vis])