training/plotting.py at main · quantling/training · GitHub

1
2
3
4
5
6
7
8
9
10
11
12
13
14
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
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
import numpy as np
import matplotlib.pyplot as plt
from matplotlib.lines import Line2D

def plot_cp(cp, file_name):
    fig = plt.figure(figsize=(10, 10))
    ax1 = fig.add_axes([0.1, 0.65, 0.8, 0.3], ylim=(-3, 3))
    ax2 = fig.add_axes([0.1, 0.35, 0.8, 0.3], xticklabels=[], sharex=ax1, sharey=ax1)
    ax3 = fig.add_axes([0.1, 0.05, 0.8, 0.3], sharex=ax1, sharey=ax1)

    for ii in range(10):
        ax1.plot(cp[:, ii], label=f'param{ii:0d}')
    ax1.legend()
    for ii in range(10, 20):
        ax2.plot(cp[:, ii], label=f'param{ii:0d}')
    ax2.legend()
    for ii in range(20, 30):
        ax3.plot(cp[:, ii], label=f'param{ii:0d}')
    ax3.legend()
    fig.savefig(file_name, dpi=300)
    plt.close('all')


def plot_mel(mel, file_name):
    fig = plt.figure(figsize=(10, 6))
    plt.imshow(mel.T, aspect='equal', vmin=-5, vmax=20)
    fig.savefig(file_name, dpi=300)
    plt.close('all')
    plt.show()

def plot_loss(res_train, res_valid, loss_type = "MSE"):
    fig, ax = plt.subplots(figsize=(15, 10), facecolor="white")
    tmp = res_train.groupby('epoch')['loss'].mean()
    plt.semilogy(np.array(tmp.index), np.array(tmp), c='C0', lw=5, label='mean train loss')
    del tmp
    tmp = res_valid.groupby('epoch')['loss'].mean()
    plt.semilogy(np.array(tmp.index), np.array(tmp), c='C1', lw=5, label='mean validation loss')
    ax.set_xlabel('Epoch', fontsize=20, labelpad=20)
    ax.set_ylabel('Loss (%s)' % str(loss_type), fontsize=20, labelpad=20)
    ax.tick_params(axis='both', labelsize=15)
    ax.legend(fontsize=15)

def plot_sublosses(res_train, res_valid):
    fig, ax = plt.subplots(figsize=(15, 10), facecolor="white")
    columns = list(res_train.columns)
    colors = ["C%d" % i for i in range(10)]

    i = 0
    for col in columns:
        if "loss" in col:
            color = colors[i]
            tmp = res_train.groupby('epoch')[col].mean()
            plt.semilogy(np.array(tmp.index), np.array(tmp), c=color, lw=5, label='%s' % str(col))
            del tmp
            tmp = res_valid.groupby('epoch')[col].mean()
            plt.semilogy(np.array(tmp.index), np.array(tmp), c=color, ls = "dotted", lw=5)
            i += 1
    ax.set_xlabel('Epoch', fontsize=20, labelpad=20)
    ax.set_ylabel('Loss' , fontsize=20, labelpad=20)
    ax.tick_params(axis='both', labelsize=15)
    leg = ax.legend(fontsize=15, bbox_to_anchor=[1.0, 1],frameon = False)
    ax.add_artist(leg)

    legend_elements = [Line2D([0], [0], color='black', lw=5, label='Training'),
                       Line2D([0], [0], color='black', lw=5, linestyle="dotted", label='Validation')]

    ax.legend(handles = legend_elements,fontsize=15, bbox_to_anchor=[.78, 1], frameon = False)


def plot_cp_predictions(prediction, target,type = "vocal tract",title=""):
    assert type in ["vocal tract", "glottis"]
    # vocal tract cps
    plt.figure(figsize=(15, 8), facecolor="white")
    # plt.plot(target_cp[:,:19], linewidth = 3 )
    if type == "vocal tract":
        plt.plot(target[:, :19], linewidth=5, alpha=0.5)
        plt.gca().set_prop_cycle(None)
        # plt.plot(target_prediction[:len(target_cp),:19],linestyle = "dotted",linewidth = 3)
        plt.plot(prediction[:len(target), :19], linestyle="dashed", linewidth=3)
        title = "Vocal Tract CPs: "+ title


    elif type == "glottis":
        plt.plot(target[:, 19:], linewidth=5,alpha=0.5)
        plt.gca().set_prop_cycle(None)
        # plt.plot(target_prediction[:len(target_cp),:19],linestyle = "dotted",linewidth = 3)
        plt.plot(prediction[:len(target), 19:], linestyle="dashed", linewidth=3)

        title = "Glottis CPs: " + title

    legend_elements = [Line2D([0], [0], color='black', lw=3, label='True CP'),
                       Line2D([0], [0], color='black', lw=3, linestyle="dotted", label='Predicted CP')]

    plt.legend(handles=legend_elements, loc='lower right', fontsize=15)
    plt.title(title, fontsize=18, pad=10)
    plt.show()


def plot_mel_predictions(target,prediction, title=""):
    fig, ax = plt.subplots(figsize=(15, 8), ncols=1, nrows=2, sharex=True, sharey=True)
    fig.set_facecolor("white")

    ax[0].set_title(title, fontsize=18, pad=20)
    img1 = target
    img2 = prediction

    ax[0].imshow(img1.T, origin='lower')
    ax[0].set_ylabel('Mel Channel', fontsize=15, rotation=270, labelpad=20)
    ax[0].yaxis.tick_right()
    ax[0].yaxis.set_label_position("right")
    ax[1].imshow(img2.T, origin='lower')
    ax[1].set_ylabel('Mel Channel', fontsize=15, rotation=270, labelpad=20)
    ax[1].yaxis.tick_right()
    ax[1].yaxis.set_label_position("right")

    ax[1].set_xlabel('Time', fontsize=15, labelpad=20)

    ax[0].text(-0.075, 0.5, "Target", fontsize=18, va="center",rotation=90,
             transform=ax[0].transAxes)
    ax[1].text(-0.075, 0.5, "Prediction", fontsize=18, va="center", rotation=90,
               transform=ax[1].transAxes)
    plt.show()