metadata
tags:
- ML-Agents-SnowballTarget
- ppo
- deep-reinforcement-learning
- reinforcement-learning
- ml-agents
model-index:
- name: PPO
results:
- task:
type: reinforcement-learning
name: reinforcement-learning
dataset:
name: ML-Agents-SnowballTarget
type: ML-Agents-SnowballTarget
metrics:
- type: mean_reward
value: 26.02 +/- 2.14
name: mean_reward
verified: false
PPO Agent playing ML-Agents-SnowballTarget
This is a trained model of a PPO agent playing ML-Agents-SnowballTarget using Unity ML-Agents.
Usage
Download model and play it:
from mlagents_envs.environment import UnityEnvironment
from mlagents_envs.base_env import ActionTuple
import numpy as np
# Load the environment
env = UnityEnvironment(file_name="path/to/SnowballTarget")
# Reset the environment
env.reset()
behavior_names = list(env.behavior_specs)
spec = env.behavior_specs[behavior_names[0]]
# Load your trained model
# (You'll need to implement model loading based on your training framework)
# Run the environment
decision_steps, terminal_steps = env.get_steps(behavior_names[0])
while True:
# Get observations
obs = decision_steps.obs[0]
# Get action from your model
# action = your_model.predict(obs)
# Step the environment
action_tuple = ActionTuple(discrete=np.array([[action]]))
env.set_actions(behavior_names[0], action_tuple)
env.step()
decision_steps, terminal_steps = env.get_steps(behavior_names[0])
if len(terminal_steps) > 0:
break
env.close()