File size: 3,292 Bytes
00ff826 | 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 | """Demon Attack training cross-task evaluation companion figure."""
from __future__ import annotations
import csv
from pathlib import Path
import sys
import matplotlib
matplotlib.use("Agg")
import matplotlib.pyplot as plt
from matplotlib.lines import Line2D
from matplotlib.ticker import FuncFormatter
HERE = Path(__file__).resolve().parent
PAPER = HERE.parents[1]
sys.path.insert(0, str(PAPER))
import paper_colors
TASKS = ("Demon Attack", "Asterix", "Atlantis", "AirRaid")
SERIES = (
("zero_latency_train", "Zero-latency training", paper_colors.GRAY),
("mixed_latency_train", "Mixed-latency training", paper_colors.PURPLE),
)
def read_rows() -> dict[str, list[dict[str, float]]]:
by_task: dict[str, list[dict[str, float]]] = {task: [] for task in TASKS}
with (HERE / "data.csv").open(newline="", encoding="utf-8") as stream:
for row in csv.DictReader(stream):
by_task[row["task"]].append({
"eval_latency": float(row["eval_latency"]),
"zero_latency_train": float(row["zero_latency_train"]),
"mixed_latency_train": float(row["mixed_latency_train"]),
})
for rows in by_task.values():
rows.sort(key=lambda item: item["eval_latency"])
return by_task
def compact_int(value: float, _position: int) -> str:
if abs(value) >= 1000:
return f"{value / 1000:g}k"
return f"{value:g}"
def main() -> None:
plt.style.use(PAPER / "plot_style.mplstyle")
plt.rcParams.update(paper_colors.RC_COLORS)
data = read_rows()
fig, axes = plt.subplots(1, 4, figsize=(7.2, 2.25), sharex=True)
fig.subplots_adjust(left=0.075, right=0.995, top=0.76, bottom=0.24, wspace=0.32)
for ax, task in zip(axes, TASKS):
rows = data[task]
xs = [row["eval_latency"] for row in rows]
max_value = max(row[key] for row in rows for key, _label, _color in SERIES)
ymin = 0
ymax = max_value * 1.18
for key, label, color in SERIES:
ys = [row[key] for row in rows]
ax.plot(
xs,
ys,
marker="o",
markersize=3.1,
linewidth=1.25,
color=color,
label=label,
)
ax.set_title(task, fontsize=8.4, weight="bold", pad=6)
ax.set_xlim(-0.2, 4.2)
ax.set_ylim(ymin, ymax)
ax.set_xticks((0, 2, 4))
ax.yaxis.set_major_formatter(FuncFormatter(compact_int))
ax.tick_params(axis="both", labelsize=6.5)
ax.grid(True, axis="y")
ax.grid(False, axis="x")
axes[0].set_ylabel("Mean return", fontsize=7.2)
fig.supxlabel("Evaluation latency (raw frames)", fontsize=7.2, y=0.07)
handles = [
Line2D([0], [0], color=color, marker="o", markersize=3.2, linewidth=1.25, label=label)
for _key, label, color in SERIES
]
fig.legend(
handles=handles,
loc="upper center",
bbox_to_anchor=(0.52, 0.995),
ncols=2,
columnspacing=1.6,
handlelength=1.8,
fontsize=6.5,
)
for extension in ("pdf", "png"):
fig.savefig(HERE / f"fig.{extension}", dpi=300, bbox_inches="tight", pad_inches=0.025)
plt.close(fig)
if __name__ == "__main__":
main()
|