Skip to content

history

Training History Plotting API

This module provides utility functions to plot training history metrics.

Functions:

Functions

plot_history_metrics 

plot_history_metrics(history: dict[str, list[float]], metrics: list[str], save_path: Path | None = None, include_val: bool = True, figsize: tuple[int, int] = (9, 5), colors: tuple[str | tuple[str, str]] = ('blue', 'orange'), stack: bool = False, title: str | None = None, **kwargs) -> tuple[plt.Figure, plt.Axes]

Plot training history metrics returned by model.fit.

Parameters:

  • history (dict[str, list[float]]) –

    Training history

  • metrics (list[str]) –

    Metrics to plot

  • save_path (Path | None, default: None ) –

    Path to save plot. Defaults to None.

  • include_val (bool, default: True ) –

    Include validation metrics. Defaults to True.

  • figsize (tuple[int, int], default: (9, 5) ) –

    Figure size. Defaults to (9, 5).

  • colors (tuple[str | tuple[str, str]], default: ('blue', 'orange') ) –

    Colors for train and val. Defaults to ("blue", "orange").

  • stack (bool, default: False ) –

    Stack metrics. Defaults to False.

  • title (str | None, default: None ) –

    Title for plot. Defaults

Returns:

  • tuple[Figure, Axes] –

    tuple[plt.Figure, plt.Axes]: Figure and axes handles

Example: ```python

history = dict( loss=[0.1, 0.2, 0.3, 0.4], accuracy=[0.9, 0.8, 0.7, 0.6], val_loss=[0.1, 0.2, 0.3, 0.4], val_accuracy=[0.9, 0.8, 0.7, 0.6], )

fig, ax = nse.plotting.plot_history_metrics(history, metrics=["loss", "accuracy"], include_val=True, stack=False)

Source code in neuralspot_edge/plotting/history.py 
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
def plot_history_metrics(
    history: dict[str, list[float]],
    metrics: list[str],
    save_path: Path | None = None,
    include_val: bool = True,
    figsize: tuple[int, int] = (9, 5),
    colors: tuple[str | tuple[str, str]] = ("blue", "orange"),
    stack: bool = False,
    title: str | None = None,
    **kwargs,
) -> tuple[plt.Figure, plt.Axes]:
    """Plot training history metrics returned by model.fit.

    Args:
        history (dict[str, list[float]]): Training history
        metrics (list[str]): Metrics to plot
        save_path (Path|None, optional): Path to save plot. Defaults to None.
        include_val (bool, optional): Include validation metrics. Defaults to True.
        figsize (tuple[int, int], optional): Figure size. Defaults to (9, 5).
        colors (tuple[str | tuple[str, str]], optional): Colors for train and val. Defaults to ("blue", "orange").
        stack (bool, optional): Stack metrics. Defaults to False.
        title (str|None, optional): Title for plot. Defaults

    Returns:
        tuple[plt.Figure, plt.Axes]: Figure and axes handles

    Example:
    ```python

    history = dict(
        loss=[0.1, 0.2, 0.3, 0.4],
        accuracy=[0.9, 0.8, 0.7, 0.6],
        val_loss=[0.1, 0.2, 0.3, 0.4],
        val_accuracy=[0.9, 0.8, 0.7, 0.6],
    )

    fig, ax = nse.plotting.plot_history_metrics(history, metrics=["loss", "accuracy"], include_val=True, stack=False)

    """
    num_axes = len(metrics) if stack else 1
    fig, ax = plt.subplots(num_axes, 1, figsize=figsize, **kwargs)
    epochs = range(1, len(history["loss"]) + 1)
    for i, metric in enumerate(metrics):
        met_ax = ax[i] if stack else ax
        if len(colors) == len(metrics):
            primary_color, secondary_color = colors[i] if isinstance(colors[i], tuple) else (colors[i], colors[i])
        else:
            primary_color, secondary_color = colors if isinstance(colors, tuple) else (colors, colors)
        met_ax.plot(epochs, history[metric], color=primary_color, label="Train" if stack else metric, linestyle="--")

        if include_val:
            met_ax.plot(
                epochs,
                history[f"val_{metric}"],
                color=secondary_color,
                label="Validation" if stack else f"val_{metric}",
            )

        if stack:
            met_ax.set_ylabel(metric)
    # END FOR

    # Set x-axis label
    if num_axes > 1:
        ax[-1].set_xlabel("Epoch")
    else:
        ax.set_xlabel("Epoch")

    # Show legend if not stacked
    if not stack:
        fig.legend()

    # Set title
    if title:
        fig.suptitle(title)

    if save_path:
        fig.savefig(save_path)

    return fig, ax