-
Notifications
You must be signed in to change notification settings - Fork 0
/
plot.py
70 lines (58 loc) · 2.43 KB
/
plot.py
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
#%% Imports
import os
import numpy as np
import glob
import tensorflow as tf
import matplotlib.pyplot as plt
#%% Inputs
log_dir = "runs"
print(os.path.join(log_dir,"*"))
path2eventsFolders = glob.glob(os.path.join(log_dir,"*"))
path2eventsFolder = path2eventsFolders[-1]
print(path2eventsFolder)
#%% Read in the results
path2eventsFile = os.path.join(path2eventsFolder, "events.out.*")
eventsFile = glob.glob(path2eventsFile)[0]
print('file',eventsFile)
#eventsFile = "runs/2022-11-09_15-22-12_SAC_HalfCheetah-v2_Gaussian_autotune/events*"
eventsFile = glob.glob("runs/2022-12-12_12-15-15_SAC_drone_2D_Gaussian_autotune/events.out.*")[0]
def get_tf_results(file, tagName):
"""
requires tensorflow==2.10.0
"""
output = []
for e in tf.compat.v1.train.summary_iterator(file):
# print(e)
for v in e.summary.value:
if v.tag == tagName:
output.append(v.simple_value)
return np.array(output)
avg_critic_loss_per_episode1 = get_tf_results(eventsFile, 'loss/critic_1')
avg_critic_loss_per_episode2 = get_tf_results(eventsFile, 'loss/critic_2')
#avg_actor_loss_per_episode = get_tf_results(eventsFile, 'loss/entropy_loss') #why am i plotting entropy loss as actor loss? shouldn't this be policy loss??
avg_actor_loss_per_episode = get_tf_results(eventsFile, 'loss/policy') #why am i plotting entropy loss as actor loss? shouldn't this be policy loss??
avg_reward_per_episode = get_tf_results(eventsFile, 'avg_reward/test')
#%% Plot results
# Number of episodes
episodes = np.arange(len(avg_actor_loss_per_episode))
# Create a figure for the critic loss
plt.figure(figsize=(8,8))
episodes = np.arange(len(avg_critic_loss_per_episode1))
plt.plot(np.arange(len(avg_critic_loss_per_episode1)), avg_critic_loss_per_episode1,'g')
plt.plot(np.arange(len(avg_critic_loss_per_episode2)), avg_critic_loss_per_episode2,'g')
plt.xlabel('Episode', fontsize=20)
plt.ylabel('Critic Loss', fontsize=20)
plt.show()
# Create a figure for the actor loss
plt.figure(figsize=(8,8))
plt.plot(np.arange(len(avg_actor_loss_per_episode)), avg_actor_loss_per_episode,'r')
plt.xlabel('Episode', fontsize=20)
plt.ylabel('Actor Loss', fontsize=20)
plt.show()
print('rewards', avg_reward_per_episode)
# Create a figure for the average reward
plt.figure(figsize=(8,8))
plt.plot(np.arange(len(avg_reward_per_episode)), avg_reward_per_episode,'b')
plt.xlabel('Episode', fontsize=20)
plt.ylabel('Average Reward', fontsize=20)
plt.show()