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
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
|
#!/usr/bin/env python3
"""Plot the P5 physical-grid classification crossover from committed JSON."""
from __future__ import annotations
import argparse
import json
from pathlib import Path
import matplotlib.pyplot as plt
import numpy as np
METHOD_STYLE = {
"raw": {
"label": "Raw local update",
"color": "#666666",
"marker": "x",
"linestyle": ":",
},
"constant": {
"label": "Constant calibration",
"color": "#E69F00",
"marker": "^",
"linestyle": "--",
},
"sdil": {
"label": "SDIL",
"color": "#0072B2",
"marker": "o",
"linestyle": "-",
},
"overclamp": {
"label": "Overclamping",
"color": "#222222",
"marker": "D",
"linestyle": "--",
},
"overclamp_sdil": {
"label": "Overclamping + SDIL",
"color": "#009E73",
"marker": "s",
"linestyle": "-",
},
}
def parse_args() -> argparse.Namespace:
parser = argparse.ArgumentParser()
parser.add_argument(
"--input", type=Path,
default=Path("results/physical_bias/p5_full_grid_bias_crossover.json"))
parser.add_argument(
"--output-prefix", type=Path,
default=Path("results/figs/physical_grid_p5_full"))
return parser.parse_args()
def values_by_diameter(records: list[dict], method: str) -> tuple[np.ndarray, list[np.ndarray]]:
diameters = np.asarray(sorted({
record["input_diameter_v"] for record in records
}))
values = []
for diameter in diameters:
values.append(np.asarray([
100.0 * record["methods"][method]["classification_error"]
for record in records
if record["input_diameter_v"] == diameter
]))
return 1_000.0 * diameters, values
def draw_panel(
axis: plt.Axes,
records: list[dict],
methods: tuple[str, ...],
title: str,
) -> None:
offsets = np.linspace(-3.0, 3.0, len(methods))
for method_index, method in enumerate(methods):
diameters, values = values_by_diameter(records, method)
style = METHOD_STYLE[method]
for diameter, trial_values in zip(diameters, values):
trial_jitter = np.linspace(-1.5, 1.5, len(trial_values))
axis.scatter(
diameter + offsets[method_index] + trial_jitter,
trial_values,
color=style["color"],
marker=style["marker"],
s=18,
alpha=0.28,
linewidths=0.8,
zorder=2,
)
means = np.asarray([np.mean(trials) for trials in values])
axis.plot(
diameters,
means,
label=style["label"],
color=style["color"],
marker=style["marker"],
linestyle=style["linestyle"],
linewidth=1.8,
markersize=5.5,
markerfacecolor=(
"white" if method in {"sdil", "overclamp_sdil"}
else style["color"]),
markeredgewidth=1.2,
zorder=3,
)
axis.set_title(title, loc="left", fontsize=10, fontweight="bold")
axis.set_xlabel("Input diameter (mV)")
axis.set_xticks(diameters)
axis.set_ylim(-2.0, 80.0)
axis.set_yticks((0, 20, 40, 60, 80))
axis.grid(axis="y", color="#D9D9D9", linewidth=0.7, zorder=0)
axis.spines[["top", "right"]].set_visible(False)
axis.legend(frameon=False, fontsize=8.5, loc="upper right")
def main() -> None:
args = parse_args()
report = json.loads(args.input.read_text())
records = report["records"]
plt.rcParams.update({
"font.family": "DejaVu Sans",
"font.size": 9,
"axes.labelsize": 9,
"xtick.labelsize": 8,
"ytick.labelsize": 8,
"svg.fonttype": "none",
"pdf.fonttype": 42,
})
figure, axes = plt.subplots(
1, 2, figsize=(7.1, 2.75), sharey=True, constrained_layout=True)
draw_panel(
axes[0], records, ("raw", "constant", "sdil"),
"(a) Standard local learning")
draw_panel(
axes[1], records, ("overclamp", "overclamp_sdil"),
"(b) Composition with overclamping")
axes[0].set_ylabel("Classification error (%)")
diameters = sorted({record["input_diameter_v"] for record in records})
trials_per_diameter = sum(
record["input_diameter_v"] == diameters[0] for record in records)
axes[1].text(
0.98, 0.62, f"{trials_per_diameter} trials per input diameter",
transform=axes[1].transAxes, ha="right", va="top",
fontsize=8, color="#666666")
args.output_prefix.parent.mkdir(parents=True, exist_ok=True)
figure.savefig(args.output_prefix.with_suffix(".svg"))
figure.savefig(args.output_prefix.with_suffix(".pdf"))
figure.savefig(args.output_prefix.with_suffix(".png"), dpi=300)
plt.close(figure)
print(f"wrote {args.output_prefix}.{{svg,pdf,png}}")
if __name__ == "__main__":
main()
|
