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

feat: Improve 2D visualization with Voronoi Cells

Browse Source
This commit is contained in:
Alexander Engelsberger 2021-10-15 13:01:01 +02:00
parent 967953442b
commit d1985571b3
No known key found for this signature in database
GPG Key ID: BE3F5909FF0D83E3
9 changed files with 109 additions and 238 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

10
examples/cbc_iris.py
View File

@ -38,10 +38,12 @@ if __name__ == "__main__":
)
# Callbacks
vis = pt.models.VisCBC2D(data=train_ds,
title="CBC Iris Example",
resolution=100,
axis_off=True)
vis = pt.models.Visualize2DVoronoiCallback(
data=train_ds,
title="CBC Iris Example",
resolution=100,
axis_off=True,
)
# Setup trainer
trainer = pl.Trainer.from_argparse_args(

12
examples/glvq_iris.py
View File

@ -3,7 +3,7 @@
import argparse
import prototorch as pt
import prototorch.models.expanded
import prototorch.models.clcc
import pytorch_lightning as pl
import torch
from torch.optim.lr_scheduler import ExponentialLR
@ -30,7 +30,7 @@ if __name__ == "__main__":
)
# Initialize the model
model = prototorch.models.expanded.GLVQ(
model = prototorch.models.GLVQ(
hparams,
optimizer=torch.optim.Adam,
prototypes_initializer=pt.initializers.SMCI(train_ds),
@ -42,7 +42,13 @@ if __name__ == "__main__":
model.example_input_array = torch.zeros(4, 2)
# Callbacks
vis = pt.models.VisGLVQ2D(data=train_ds)
vis = pt.models.Visualize2DVoronoiCallback(
data=train_ds,
resolution=200,
title="Example: GLVQ on Iris",
x_label="sepal length",
y_label="petal length",
)
# Setup trainer
trainer = pl.Trainer.from_argparse_args(

0
prototorch/models/clcc/__init__.py Normal file
View File

2
prototorch/models/expanded/clcc_glvq.py → prototorch/models/clcc/clcc_glvq.py
View File

@ -7,7 +7,7 @@ from prototorch.core.components import LabeledComponents
from prototorch.core.distances import euclidean_distance
from prototorch.core.initializers import AbstractComponentsInitializer, LabelsInitializer
from prototorch.core.losses import GLVQLoss
from prototorch.models.expanded.clcc_scheme import CLCCScheme
from prototorch.models.clcc.clcc_scheme import CLCCScheme
from prototorch.nn.wrappers import LambdaLayer

9
prototorch/models/expanded/clcc_scheme.py → prototorch/models/clcc/clcc_scheme.py
View File

@ -9,6 +9,7 @@ CLCC is a LVQ scheme containing 4 steps
"""
import pytorch_lightning as pl
import torch
class CLCCScheme(pl.LightningModule):
@ -36,6 +37,8 @@ class CLCCScheme(pl.LightningModule):
return comparison_tensor
def forward(self, batch):
if isinstance(batch, torch.Tensor):
batch = (batch, None)
# TODO: manage different datatypes?
components = self.components_layer()
# TODO: => Component Hook
@ -43,6 +46,12 @@ class CLCCScheme(pl.LightningModule):
# TODO: => Competition Hook
return self.inference(comparison_tensor, components)
def predict(self, batch):
"""
Alias for forward
"""
return self.forward(batch)
def loss_forward(self, batch):
# TODO: manage different datatypes?
components = self.components_layer()

15
prototorch/models/expanded/test_clcc.py → prototorch/models/clcc/test_clcc.py
View File

@ -3,12 +3,12 @@ import prototorch as pt
import pytorch_lightning as pl
import torch
from prototorch.core.initializers import SMCI, RandomNormalCompInitializer
from prototorch.models.expanded.clcc_glvq import GLVQ, GLVQhparams
from torch.utils.data import DataLoader, Dataset
from torchvision import datasets
from torchvision.transforms import Compose, Lambda, ToTensor
from prototorch.models.clcc.clcc_glvq import GLVQ, GLVQhparams
from prototorch.models.vis import Visualize2DVoronoiCallback
plt.gray()
# NEW STUFF
# ##############################################################################
# ##############################################################################
if __name__ == "__main__":
# Dataset
@ -29,7 +29,8 @@ if __name__ == "__main__":
print(model)
# Callbacks
vis = pt.models.VisGLVQ2D(data=train_ds)
vis = Visualize2DVoronoiCallback(data=train_ds, resolution=500)
# Train
trainer = pl.Trainer(callbacks=[vis], gpus=1)
trainer = pl.Trainer(callbacks=[vis], gpus=1, max_epochs=100)
trainer.fit(model, train_loader)

1
prototorch/models/expanded/__init__.py
View File

@ -1 +0,0 @@
from .glvq import GLVQ

164
prototorch/models/expanded/glvq.py
View File

@ -1,164 +0,0 @@
import pytorch_lightning as pl
import torch
import torchmetrics
from prototorch.core.competitions import WTAC, wtac
from prototorch.core.components import Components, LabeledComponents
from prototorch.core.distances import (
euclidean_distance,
lomega_distance,
omega_distance,
squared_euclidean_distance,
)
from prototorch.core.initializers import EyeTransformInitializer, LabelsInitializer
from prototorch.core.losses import GLVQLoss, lvq1_loss, lvq21_loss
from prototorch.core.pooling import stratified_min_pooling
from prototorch.core.transforms import LinearTransform
from prototorch.nn.wrappers import LambdaLayer, LossLayer
from torch.nn.parameter import Parameter
class GLVQ(pl.LightningModule):
def __init__(self, hparams, **kwargs):
super().__init__()
# Hyperparameters
self.save_hyperparameters(hparams)
# Default hparams
# TODO: Manage by an HPARAMS Object
self.hparams.setdefault("lr", 0.01)
self.hparams.setdefault("margin", 0.0)
self.hparams.setdefault("transfer_fn", "identity")
self.hparams.setdefault("transfer_beta", 10.0)
# Default config
self.optimizer = kwargs.get("optimizer", torch.optim.Adam)
self.lr_scheduler = kwargs.get("lr_scheduler", None)
self.lr_scheduler_kwargs = kwargs.get("lr_scheduler_kwargs", dict())
distance_fn = kwargs.get("distance_fn", euclidean_distance)
prototypes_initializer = kwargs.get("prototypes_initializer", None)
labels_initializer = kwargs.get("labels_initializer",
LabelsInitializer())
if prototypes_initializer is not None:
self.proto_layer = LabeledComponents(
distribution=self.hparams.distribution,
components_initializer=prototypes_initializer,
labels_initializer=labels_initializer,
)
self.distance_layer = LambdaLayer(distance_fn)
self.competition_layer = WTAC()
self.loss = GLVQLoss(
margin=self.hparams.margin,
transfer_fn=self.hparams.transfer_fn,
beta=self.hparams.transfer_beta,
)
def log_acc(self, distances, targets, tag):
preds = self.predict_from_distances(distances)
accuracy = torchmetrics.functional.accuracy(preds.int(), targets.int())
self.log(tag,
accuracy,
on_step=False,
on_epoch=True,
prog_bar=True,
logger=True)
def configure_optimizers(self):
optimizer = self.optimizer(self.parameters(), lr=self.hparams.lr)
if self.lr_scheduler is not None:
scheduler = self.lr_scheduler(optimizer,
**self.lr_scheduler_kwargs)
sch = {
"scheduler": scheduler,
"interval": "step",
} # called after each training step
return [optimizer], [sch]
else:
return optimizer
def shared_step(self, batch, batch_idx, optimizer_idx=None):
x, y = batch
out = self.compute_distances(x)
_, plabels = self.proto_layer()
loss = self.loss(out, y, plabels)
return out, loss
def training_step(self, batch, batch_idx, optimizer_idx=None):
out, train_loss = self.shared_step(batch, batch_idx, optimizer_idx)
self.log_prototype_win_ratios(out)
self.log("train_loss", train_loss)
self.log_acc(out, batch[-1], tag="train_acc")
return train_loss
def validation_step(self, batch, batch_idx):
out, val_loss = self.shared_step(batch, batch_idx)
self.log("val_loss", val_loss)
self.log_acc(out, batch[-1], tag="val_acc")
return val_loss
def test_step(self, batch, batch_idx):
out, test_loss = self.shared_step(batch, batch_idx)
self.log_acc(out, batch[-1], tag="test_acc")
return test_loss
def test_epoch_end(self, outputs):
test_loss = 0.0
for batch_loss in outputs:
test_loss += batch_loss.item()
self.log("test_loss", test_loss)
# API
def compute_distances(self, x):
protos, _ = self.proto_layer()
distances = self.distance_layer(x, protos)
return distances
def forward(self, x):
distances = self.compute_distances(x)
_, plabels = self.proto_layer()
winning = stratified_min_pooling(distances, plabels)
y_pred = torch.nn.functional.softmin(winning)
return y_pred
def predict_from_distances(self, distances):
with torch.no_grad():
_, plabels = self.proto_layer()
y_pred = self.competition_layer(distances, plabels)
return y_pred
def predict(self, x):
with torch.no_grad():
distances = self.compute_distances(x)
y_pred = self.predict_from_distances(distances)
return y_pred
@property
def prototype_labels(self):
return self.proto_layer.labels.detach().cpu()
@property
def num_classes(self):
return self.proto_layer.num_classes
@property
def num_prototypes(self):
return len(self.proto_layer.components)
@property
def prototypes(self):
return self.proto_layer.components.detach().cpu()
@property
def components(self):
"""Only an alias for the prototypes."""
return self.prototypes
# Python overwrites
def __repr__(self):
surep = super().__repr__()
indented = "".join([f"\t{line}\n" for line in surep.splitlines()])
wrapped = f"ProtoTorch Bolt(\n{indented})"
return wrapped

134
prototorch/models/vis.py
View File

@ -5,14 +5,18 @@ import pytorch_lightning as pl
import torch
import torchvision
from matplotlib import pyplot as plt
from prototorch.utils.utils import mesh2d
from prototorch.utils.utils import generate_mesh, mesh2d
from torch.utils.data import DataLoader, Dataset
COLOR_UNLABELED = 'w'
class Vis2DAbstract(pl.Callback):
def __init__(self,
data,
title="Prototype Visualization",
title=None,
x_label=None,
y_label=None,
cmap="viridis",
border=0.1,
resolution=100,
@ -44,6 +48,8 @@ class Vis2DAbstract(pl.Callback):
self.y_train = y
self.title = title
self.x_label = x_label
self.y_label = y_label
self.fig = plt.figure(self.title)
self.cmap = cmap
self.border = border
@ -56,20 +62,19 @@ class Vis2DAbstract(pl.Callback):
self.pause_time = pause_time
self.block = block
def precheck(self, trainer):
if self.show_last_only:
if trainer.current_epoch != trainer.max_epochs - 1:
return False
def show_on_current_epoch(self, trainer):
if self.show_last_only and trainer.current_epoch != trainer.max_epochs - 1:
return False
return True
def setup_ax(self, xlabel=None, ylabel=None):
def setup_ax(self):
ax = self.fig.gca()
ax.cla()
ax.set_title(self.title)
if xlabel:
ax.set_xlabel("Data dimension 1")
if ylabel:
ax.set_ylabel("Data dimension 2")
if self.x_label:
ax.set_xlabel(self.x_label)
if self.x_label:
ax.set_ylabel(self.y_label)
if self.axis_off:
ax.axis("off")
return ax
@ -116,27 +121,64 @@ class Vis2DAbstract(pl.Callback):
plt.close()
class VisGLVQ2D(Vis2DAbstract):
class Visualize2DVoronoiCallback(Vis2DAbstract):
def __init__(self, data, **kwargs):
super().__init__(data, **kwargs)
self.data_min = torch.min(self.x_train, axis=0).values
self.data_max = torch.max(self.x_train, axis=0).values
def current_span(self, proto_values):
proto_min = torch.min(proto_values, axis=0).values
proto_max = torch.max(proto_values, axis=0).values
overall_min = torch.minimum(proto_min, self.data_min)
overall_max = torch.maximum(proto_max, self.data_max)
return overall_min, overall_max
def get_voronoi_diagram(self, min, max, model):
mesh_input, (xx, yy) = generate_mesh(
min,
max,
border=self.border,
resolution=self.resolution,
device=model.device,
)
y_pred = model.predict(mesh_input)
return xx, yy, y_pred.reshape(xx.shape)
def on_epoch_end(self, trainer, pl_module):
if not self.precheck(trainer):
if not self.show_on_current_epoch(trainer):
return True
protos = pl_module.prototypes
plabels = pl_module.prototype_labels
x_train, y_train = self.x_train, self.y_train
ax = self.setup_ax(xlabel="Data dimension 1",
ylabel="Data dimension 2")
self.plot_data(ax, x_train, y_train)
self.plot_protos(ax, protos, plabels)
x = np.vstack((x_train, protos))
mesh_input, xx, yy = mesh2d(x,
self.border,
self.resolution,
device=pl_module.device)
mesh_input = (mesh_input, None)
y_pred = pl_module(mesh_input)
y_pred = y_pred.cpu().reshape(xx.shape)
ax.contourf(xx.cpu(), yy.cpu(), y_pred, cmap=self.cmap, alpha=0.35)
# Extract Prototypes
proto_values = pl_module.prototypes
if hasattr(pl_module, "prototype_labels"):
proto_labels = pl_module.prototype_labels
else:
proto_labels = COLOR_UNLABELED
# Calculate Voronoi Diagram
overall_min, overall_max = self.current_span(proto_values)
xx, yy, y_pred = self.get_voronoi_diagram(
overall_min,
overall_max,
pl_module,
)
ax = self.setup_ax()
ax.contourf(
xx.cpu(),
yy.cpu(),
y_pred.cpu(),
cmap=self.cmap,
alpha=0.35,
)
self.plot_data(ax, self.x_train, self.y_train)
self.plot_protos(ax, proto_values, proto_labels)
self.log_and_display(trainer, pl_module)
@ -147,7 +189,7 @@ class VisSiameseGLVQ2D(Vis2DAbstract):
self.map_protos = map_protos
def on_epoch_end(self, trainer, pl_module):
if not self.precheck(trainer):
if not self.show_on_current_epoch(trainer):
return True
protos = pl_module.prototypes
@ -185,7 +227,7 @@ class VisGMLVQ2D(Vis2DAbstract):
self.ev_proj = ev_proj
def on_epoch_end(self, trainer, pl_module):
if not self.precheck(trainer):
if not self.show_on_current_epoch(trainer):
return True
protos = pl_module.prototypes
@ -212,40 +254,16 @@ class VisGMLVQ2D(Vis2DAbstract):
self.log_and_display(trainer, pl_module)
class VisCBC2D(Vis2DAbstract):
def on_epoch_end(self, trainer, pl_module):
if not self.precheck(trainer):
return True
x_train, y_train = self.x_train, self.y_train
protos = pl_module.components
ax = self.setup_ax(xlabel="Data dimension 1",
ylabel="Data dimension 2")
self.plot_data(ax, x_train, y_train)
self.plot_protos(ax, protos, "w")
x = np.vstack((x_train, protos))
mesh_input, xx, yy = mesh2d(x, self.border, self.resolution)
_components = pl_module.components_layer._components
y_pred = pl_module.predict(
torch.Tensor(mesh_input).type_as(_components))
y_pred = y_pred.cpu().reshape(xx.shape)
ax.contourf(xx, yy, y_pred, cmap=self.cmap, alpha=0.35)
self.log_and_display(trainer, pl_module)
class VisNG2D(Vis2DAbstract):
def on_epoch_end(self, trainer, pl_module):
if not self.precheck(trainer):
if not self.show_on_current_epoch(trainer):
return True
x_train, y_train = self.x_train, self.y_train
protos = pl_module.prototypes
cmat = pl_module.topology_layer.cmat.cpu().numpy()
ax = self.setup_ax(xlabel="Data dimension 1",
ylabel="Data dimension 2")
ax = self.setup_ax()
self.plot_data(ax, x_train, y_train)
self.plot_protos(ax, protos, "w")
@ -316,7 +334,7 @@ class VisImgComp(Vis2DAbstract):
)
def on_epoch_end(self, trainer, pl_module):
if not self.precheck(trainer):
if not self.show_on_current_epoch(trainer):
return True
if self.show: