From e7446664feea65c2fe843d6a57d6b5a6f6313ba2 Mon Sep 17 00:00:00 2001
From: dcarron <daniel.carron@idiap.ch>
Date: Tue, 11 Apr 2023 14:35:37 +0200
Subject: [PATCH] Moved logistic_regression model to lightning
---
.../configs/models/logistic_regression.py | 21 ++---
src/ptbench/models/logistic_regression.py | 79 ++++++++++++++++---
2 files changed, 79 insertions(+), 21 deletions(-)
diff --git a/src/ptbench/configs/models/logistic_regression.py b/src/ptbench/configs/models/logistic_regression.py
index b93935b4..145dddd7 100644
--- a/src/ptbench/configs/models/logistic_regression.py
+++ b/src/ptbench/configs/models/logistic_regression.py
@@ -7,20 +7,23 @@
Simple feedforward network taking radiological signs in output and
predicting tuberculosis presence in output.
"""
-
+from torch import empty
from torch.nn import BCEWithLogitsLoss
-from torch.optim import Adam
-from ...models.logistic_regression import build_logistic_regression
+from ...models.logistic_regression import LogisticRegression
# config
-lr = 1e-2
-
-# model
-model = build_logistic_regression(14)
+optimizer_configs = {"lr": 1e-2}
+input_size = 14
# optimizer
-optimizer = Adam(model.parameters(), lr=lr)
+optimizer = "Adam"
# criterion
-criterion = BCEWithLogitsLoss()
+criterion = BCEWithLogitsLoss(pos_weight=empty(1))
+criterion_valid = BCEWithLogitsLoss(pos_weight=empty(1))
+
+# model
+model = LogisticRegression(
+ criterion, criterion_valid, optimizer, optimizer_configs, input_size
+)
diff --git a/src/ptbench/models/logistic_regression.py b/src/ptbench/models/logistic_regression.py
index 7e7818c7..684155b4 100644
--- a/src/ptbench/models/logistic_regression.py
+++ b/src/ptbench/models/logistic_regression.py
@@ -2,16 +2,32 @@
#
# SPDX-License-Identifier: GPL-3.0-or-later
+import pytorch_lightning as pl
import torch
import torch.nn as nn
-class LogisticRegression(nn.Module):
+class LogisticRegression(pl.LightningModule):
"""Radiological signs to Tuberculosis module."""
- def __init__(self, input_size):
+ def __init__(
+ self,
+ criterion,
+ criterion_valid,
+ optimizer,
+ optimizer_configs,
+ input_size,
+ ):
super().__init__()
- self.linear = torch.nn.Linear(input_size, 1)
+
+ self.save_hyperparameters()
+
+ self.criterion = criterion
+ self.criterion_valid = criterion_valid
+
+ self.name = "logistic_regression"
+
+ self.linear = nn.Linear(input_size, 1)
def forward(self, x):
"""
@@ -32,15 +48,54 @@ class LogisticRegression(nn.Module):
return output
+ def training_step(self, batch, batch_idx):
+ images = batch[1]
+ labels = batch[2]
+
+ # Increase label dimension if too low
+ # Allows single and multiclass usage
+ if labels.ndim == 1:
+ labels = torch.reshape(labels, (labels.shape[0], 1))
+
+ # Forward pass on the network
+ outputs = self(images)
+
+ training_loss = self.criterion(outputs, labels.double())
+
+ return {"loss": training_loss}
+
+ def validation_step(self, batch, batch_idx):
+ images = batch[1]
+ labels = batch[2]
+
+ # Increase label dimension if too low
+ # Allows single and multiclass usage
+ if labels.ndim == 1:
+ labels = torch.reshape(labels, (labels.shape[0], 1))
+
+ # data forwarding on the existing network
+ outputs = self(images)
+ validation_loss = self.criterion_valid(outputs, labels.double())
+
+ return {"validation_loss": validation_loss}
+
+ def predict_step(self, batch, batch_idx, grad_cams=False):
+ names = batch[0]
+ images = batch[1]
+
+ outputs = self(images)
+ probabilities = torch.sigmoid(outputs)
+
+ # necessary check for HED architecture that uses several outputs
+ # for loss calculation instead of just the last concatfuse block
+ if isinstance(outputs, list):
+ outputs = outputs[-1]
-def build_logistic_regression(input_size):
- """Build logistic regression module.
+ return names[0], torch.flatten(probabilities), torch.flatten(batch[2])
- Returns
- -------
+ def configure_optimizers(self):
+ optimizer = getattr(torch.optim, self.hparams.optimizer)(
+ self.parameters(), **self.hparams.optimizer_configs
+ )
- module : :py:class:`torch.nn.Module`
- """
- model = LogisticRegression(input_size)
- model.name = "logistic_regression"
- return model
+ return optimizer
--
GitLab