Plotter

Plotter generates diagnostic plots (loss curve, learning-rate schedule) during and after training. All figures are saved as PNG files so they work in headless / HPC environments.

Plotter

nnfe.plotter.Plotter

Generates and saves diagnostic plots for an NNFE training run.

Parameters:

Name	Type	Description	Default
config		:class:~nnfe.nnfe_config.PlotterConfig instance.	required
save_dir		Directory path for saving plot files. Should be a :class:pathlib.Path.	None

Attributes:

Name	Type	Description
config		:class:~nnfe.nnfe_config.PlotterConfig controlling which plots are produced.
save_dir		Directory where PNG files are written. If None saving is skipped (no plots are generated).

plot_loss(loss_vec)

Plot the training loss on a log scale and save to loss.png.

Parameters:

Name	Type	Description	Default
loss_vec		1-D NumPy array of per-epoch loss values.	required

plot_learning_rate(scheduler, epochs)

Plot the learning-rate schedule on a log scale and save to lr.png.

Parameters:

Name	Type	Description	Default
scheduler		An Optax schedule callable (step: int) -> float.	required
epochs		Total number of training epochs; determines the x-axis range.	required

plot_grad()

Placeholder for gradient-norm diagnostics (not yet implemented).