File size: 4,532 Bytes
337a98d | 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 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 | # coding=utf-8
# Copyright 2024 The Google Research Authors.
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
"""Example script for generating plots like Figures 3 and 4 of the paper."""
import dataclasses
import os
import pathlib
from typing import List
from absl import app
from absl import flags
import matplotlib.pyplot as plt
import numpy as np
import pandas as pd
from tensorboard.backend.event_processing import event_accumulator
_PLOT_DIR = flags.DEFINE_string(
"plot_dir",
"/tmp/xirl/plots",
"Directory wherein to store plots.")
_MATPLOTLIB_SIZE = flags.DEFINE_integer(
"size",
15,
"Size multiplier for matplotlib figure.")
_STD_DEV_FRAC = flags.DEFINE_float(
"std_dev_frac",
0.5,
"+/- standard deviation fraction.")
_MATPLOTLIB_DPI = flags.DEFINE_integer(
"dpi",
100,
"Dots per inch for PNG plots.")
def update_plotting_params(size):
plt.rcParams.update({
"legend.fontsize": "large",
"axes.titlesize": size,
"axes.labelsize": size,
"xtick.labelsize": size,
"ytick.labelsize": size,
})
def minimum_truncate_array_list(arrs):
min_len = arrs[0].shape[0]
for arr in arrs[1:]:
if arr.shape[0] < min_len:
min_len = arr.shape[0]
return [arr[:min_len] for arr in arrs]
@dataclasses.dataclass
class Experiment:
"""Parses experiment data and computes statistics across seeds."""
path: str
name: str
color: str
linestyle: str
def __post_init__(self):
self.path = pathlib.Path(self.path)
if not self.path.exists():
raise ValueError(f"{self.path} does not exist.")
# Find seed directories.
subdirs = [f for f in self.path.iterdir() if f.is_dir()]
# cd into their respective log subdirs.
logdirs = [subdir / "tb" for subdir in subdirs]
# Read Tensorboard logs.
logfiles = [list(logdir.iterdir())[0] for logdir in logdirs]
data = []
for logfile in logfiles:
ea = event_accumulator.EventAccumulator(str(logfile))
ea.Reload()
df = pd.DataFrame(ea.Scalars("evaluation/average_eval_scores"))
arr = df[["step", "value"]].to_numpy()
data.append(arr)
self.data = minimum_truncate_array_list(data)
@property
def mean(self):
return np.mean(self.data, axis=0)
@property
def std_dev(self):
return np.std(self.data, axis=0)
def cross_shortstick(savename):
"""Aggregates returns across experiment seeds and generates a figure."""
# Note: Append baselines or other methods to this list.
experiments = [
Experiment(
# Note: replace with an actual experiment path.
path="/PATH/TO/AN/EXPERIMENT/HERE/",
# Note: You can customize the below attributes to your liking.
name="XIRL",
color="tab:red",
linestyle="dashdot",
),
]
_, ax = plt.subplots(1, constrained_layout=True)
for experiment in experiments:
return_mean = experiment.mean
return_stddev = experiment.std_dev
return_mean_x = return_mean[:, 0] / 1_000
return_mean_y = return_mean[:, 1]
ax.plot(
return_mean_x,
return_mean_y,
lw=2,
label=experiment.name,
color=experiment.color,
linestyle=experiment.linestyle,
)
ax.fill_between(
return_mean_x,
return_mean_y + _STD_DEV_FRAC.value * return_stddev[:, 1],
return_mean_y - _STD_DEV_FRAC.value * return_stddev[:, 1],
alpha=0.2,
color=experiment.color,
)
ax.set_xlabel("Steps (thousands)")
ax.set_ylabel("Success Rate")
ax.set_title("short-stick")
ax.legend(loc="lower right")
ax.grid(linestyle="--", linewidth=0.5)
plt.savefig(f"{_PLOT_DIR.value}/{savename}.pdf", format="pdf")
plt.savefig(
f"{_PLOT_DIR.value}/{savename}.png",
format="png",
dpi=_MATPLOTLIB_DPI.value,
)
plt.close()
def main(_):
os.makedirs(_PLOT_DIR.value, exist_ok=True)
update_plotting_params(_MATPLOTLIB_SIZE.value)
cross_shortstick("cross_embodiment_shortstick")
if __name__ == "__main__":
app.run(main)
|