Spaces:
Running
Running
File size: 4,456 Bytes
ba246bb |
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 |
import matplotlib.pyplot as plt
from numpy import ndarray
import seaborn as sns
import numpy as np
from scipy.stats import gaussian_kde
def plot_test_results(rewards: ndarray, wins: ndarray, steps: ndarray, normalized_return: ndarray) -> None:
"""
Plot the rewards obtained on the test data, the number of times the agent wins, and the length (number of time steps)
of each episode.
:param rewards: list of rewards obtained on the test data
:param wins: list of wins obtained on the test data
:param steps: list of steps of each episode obtained on the test data
:param normalized_return: list of normalized return obtained on the test data
"""
nb_episodes = len(rewards)
cat = [i for i in range(nb_episodes)]
plt.figure(figsize=(10, 6))
plt.subplot(4, 1, 1) # 4 lignes, 1 colonne, graphique 1
plt.bar(cat, rewards, label='avg_rewards')
plt.title('Average rewards on test data')
plt.legend()
# Ajouter le deuxième sous-graphe
plt.subplot(4, 1, 2) # 4 lignes, 1 colonne, graphique 2
plt.bar(cat, wins, label='avg_wins', color='orange')
plt.title('Average wins on test data')
plt.legend()
# Ajouter le troisième sous-graphe
plt.subplot(4, 1, 3) # 4 lignes, 1 colonne, graphique 3
plt.bar(cat, steps, label='avg_steps', color='green')
plt.title('Average length of episodes on test data')
plt.legend()
# Ajouter le quatrième sous-graphe
plt.subplot(4, 1, 4) # 4 lignes, 1 colonne, graphique 4
plt.bar(cat, normalized_return, label='avg_normalized_return', color='green')
plt.title('Average normalized return obtained on test data')
plt.legend()
# Ajuster l'espacement entre les sous-graphes
plt.tight_layout()
# Afficher les graphiques
plt.show()
def plot_density(normalized_return: ndarray):
# Figure
plt.figure(figsize=(15, 4))
# --- 1. Histogramme brut ---
plt.subplot(1, 3, 1)
plt.hist(normalized_return, bins=10, color="green", edgecolor="black")
plt.title("Histogramme brut")
plt.xlabel("Retour normalisé")
plt.ylabel("Nombre de maillages")
# --- 2. Histogramme + densité normalisée ---
plt.subplot(1, 3, 2)
sns.histplot(normalized_return, bins=10, color="green", edgecolor="black", stat="density", kde=False,
label="Histogramme (densité)")
sns.kdeplot(normalized_return, color="red", linewidth=2, label="Densité KDE", clip=(0,1))
plt.title("Histogramme + densité normalisée")
plt.xlabel("Retour normalisé")
plt.ylabel("Densité")
plt.legend()
# --- 3. Histogramme + densité mise à l’échelle ---
plt.subplot(1, 3, 3)
counts, bins, _ = plt.hist(normalized_return, bins=10, color="green", edgecolor="black", label="Histogramme")
# KDE avec mise à l'échelle
x = np.linspace(0, 1, 200)
kde = gaussian_kde(normalized_return)
kde_scaled = kde(x) * len(normalized_return) * (bins[1] - bins[0])
plt.plot(x, kde_scaled, color="red", linewidth=2, label="Densité mise à l’échelle")
plt.title("Histogramme + densité à l’échelle")
plt.xlabel("Retour normalisé")
plt.ylabel("Nombre de maillages")
plt.xlim(0, 1)
plt.legend()
plt.tight_layout()
plt.show()
def plot_training_results(rewards: ndarray, wins: ndarray, steps: ndarray) -> None:
"""
Plot the rewards obtained during training, the number of times the agent wins, and the length (number of time steps)
of each episode.
:param rewards: list of rewards obtained during training
:param wins: list of wins obtained during training
:param steps: list of steps obtained during training
"""
nb_episodes = len(rewards)
real_runs = 1
plt.figure()
plt.plot(rewards)
plt.xlabel('Episodes')
plt.ylabel('Rewards')
plt.title('Learning Rewards, nb_episodes={}'.format(nb_episodes) + ', nb_runs={}'.format(real_runs))
plt.legend(loc="best")
plt.show()
plt.figure()
plt.plot(wins)
plt.xlabel('Episodes')
plt.ylabel('Wins')
plt.title('Learning Wins, nb_episodes={}'.format(nb_episodes) + ', nb_runs={}'.format(real_runs))
plt.legend(loc="best")
plt.show()
plt.figure()
plt.plot(steps)
plt.xlabel('Episodes')
plt.ylabel('Number of steps per episode')
plt.title('Learning steps, nb_episodes={}'.format(nb_episodes) + ', nb_runs={}'.format(real_runs))
plt.legend(loc="best")
plt.show() |