dispatcher / plot_rewards.py
garg-tejas
rebrand: Dead Air -> DispatchR across all code, docs, and metadata
e687796
Raw History Blame Contribute Delete
1.66 kB
"""Plot training reward curves."""
import argparse
import json
from pathlib import Path
import matplotlib.pyplot as plt
import numpy as np
def plot_rewards(rewards: list, output: str = "training_reward.png", title: str = "DispatchR Training Reward"):
"""Plot reward curve with moving average."""
plt.figure(figsize=(12, 5))
# Raw rewards
plt.subplot(1, 2, 1)
plt.plot(rewards, alpha=0.5, label="Episode Reward")
if len(rewards) >= 10:
ma = np.convolve(rewards, np.ones(10)/10, mode="valid")
plt.plot(range(9, len(rewards)), ma, color="red", label="10-ep Moving Avg")
plt.xlabel("Episode")
plt.ylabel("Reward")
plt.title(title)
plt.legend()
plt.grid(True)
# Histogram
plt.subplot(1, 2, 2)
plt.hist(rewards, bins=20, edgecolor="black")
plt.xlabel("Reward")
plt.ylabel("Frequency")
plt.title("Reward Distribution")
plt.grid(True)
plt.tight_layout()
plt.savefig(output)
plt.close()
print(f"Saved plot to {output}")
def main():
parser = argparse.ArgumentParser()
parser.add_argument("--input", type=str, default="outputs/rewards.json")
parser.add_argument("--output", type=str, default="assets/training_reward.png")
args = parser.parse_args()
with open(args.input) as f:
data = json.load(f)
rewards = data.get("rewards", [])
print(f"Loaded {len(rewards)} episodes")
print(f"Mean: {np.mean(rewards):.3f}, Max: {np.max(rewards):.3f}, Min: {np.min(rewards):.3f}")
Path(args.output).parent.mkdir(parents=True, exist_ok=True)
plot_rewards(rewards, output=args.output)
if __name__ == "__main__":
main()