Skip to content

Adding capability for plotting loss curves #153

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Draft
namrathaurs wants to merge 3 commits into
base: main
Choose a base branch
from
Draft

Adding capability for plotting loss curves #153

Changes from all commits
Commits
Show all changes
3 commits
Select commit Hold shift + click to select a range
81765d4
initial round of development for loss curves
namrathaurs Sep 8, 2025
ce2e329
fix mismatched indentation
namrathaurs Sep 8, 2025
dcc31b7
fixing inconsistent indentation
namrathaurs Sep 8, 2025
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
55 changes: 55 additions & 0 deletions src/deepdiagnostics/plots/lossPlot.py
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -0,0 +1,55 @@
import matplotlib.pyplot as plt
from typing import Dict, List

from deepdiagnostics.plots.plot import Display

class LossPlot(Display):
def __init__(self, model=None, data=None):
...

def _data_setup(self):
raise NotImplementedError

def plot_name(self):
return "loss_curves.png"

def plot(
self,
training_history: Dict[str, List[float]],
epochs: int = None,
best_val_loss: float = None
):
try:
training_loss_data = training_history["train_loss"]
validation_loss_data = training_history["val_loss"]
except Exception as e:
raise KeyError(f"Key {e} not found in supplied training history")

if epochs is None:
if len(training_loss_data) != len(validation_loss_data):
raise ValueError("Inconsistent training history data supplied [length of train and validation losses not equal]")
epochs = len(training_loss_data)
print(f"Number of epochs determined: {epochs}")
else:
if epochs != len(training_loss_data) or epochs != len(validation_loss_data):
raise ValueError("Epochs supplied inconsistent with training history data [epochs and loss history not equal]")

if best_val_loss is None:
best_val_loss = min(validation_loss_data)
print(f"Best validation loss found: {best_val_loss}")

epochs_trained = [ x for x in range(1, epochs+1) ]
plt.plot(epochs_trained, training_loss_data, label='Training Loss')
plt.plot(epochs_trained, validation_loss_data, label='Validation Loss')
plt.axhline(y=best_val_loss, color='m', linestyle='--', label="Best Val. Loss")

plt.xlabel('Epochs')
plt.ylabel('Loss')
plt.title('Training and Validation Loss Over Epochs')

plt.legend()
plt.grid()
plt.show()
Copy link
Contributor

@voetberg voetberg Sep 10, 2025

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Two things:

  • Please use the super _finish to either save or show the plot
  • Please return the figure, axes object so the user can edit them outside of the plot method if they want


def __call__(self, **kwargs):
raise NotImplementedError("Plotting loss is not supported in pipeline mode")
Loading