gym-plotter 0.1.1

Live plotting utility for RL environments (Gym, Stable-Baselines, etc.)

gym-plotter

A lightweight live plotting utility for RL training (works with Gym, Stable-Baselines or custom loops).
Plots raw + smoothed rewards, loss and other metrics in real-time.

Features

• Live plotting during training
• Raw and smoothed (moving average) metrics
• Best score marker
• Training stats box (latest reward, avg reward)
• Auto-save final plot

Installation

pip install gym-plotter


## Usage
import gymnasium as gym
from gym_plotter import RLPlotter

# Initialize environment and live plotter
env = gym.make("CartPole-v1")
plotter = RLPlotter()

for episode in range(50):
    state, _ = env.reset()
    done = False
    total_reward = 0
    loss = 0  # Replace with your actual loss calculation

    while not done:
        action = env.action_space.sample()  # Replace with your policy action
        state, reward, terminated, truncated, info = env.step(action)
        done = terminated or truncated
        total_reward += reward

    # Log metrics after each episode
    plotter.log(reward=total_reward, loss=loss)

# Save final plot to file
plotter.save("training_plot.png")
env.close()
# Example Output

gym-plotter displays training progress in real time.
Each metric (reward, loss, etc.) appears as both raw and smoothed curves, with the best reward highlighted and a live stats box summarizing performance.