Copy over experiments

scratch/etienne/intersimple/ppo_tspeed_random.py

@@ -0,0 +1,31 @@
+# %%
+from stable_baselines3 import PPO
+from stable_baselines3.common.env_util import make_vec_env
+from intersim.envs.intersimple import IntersimpleTargetSpeedRandom
+
+model_name = "ppo_tspeed_random"
+
+# %%
+env = IntersimpleTargetSpeedRandom(target_speed=10)
+
+# %%
+model = PPO("MlpPolicy", env, verbose=1)
+model.learn(total_timesteps=250000)
+model.save(model_name)
+
+print('Done training.')
+
+del model # remove to demonstrate saving and loading
+
+# %%
+model = PPO.load(model_name)
+
+obs = env.reset()
+while True:
+    action, _states = model.predict(obs)
+    obs, rewards, done, info = env.step(action)
+    env.render(mode='post')
+    if done:
+        break
+
+env.close()