Skip to content

Use loop break instead of sys.exit() on early stop #123

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
dnddnjs merged 1 commit into from
May 17, 2026
Merged

Use loop break instead of sys.exit() on early stop #123

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
9 changes: 5 additions & 4 deletions 2-cartpole/1-dqn.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -18,7 +18,6 @@
L(theta) = ( Q_theta(s)[a] - y )^2
"""
import random
import sys
from collections import deque

import numpy as np
Expand Down Expand Up @@ -135,8 +134,11 @@ def train_model(self):
run_test_loop(env, agent.get_action)

scores = []
solved = False

for e in range(EPISODES):
if solved:
break
done = False
score = 0
state, _ = env.reset()
Expand Down Expand Up @@ -168,9 +170,8 @@ def train_model(self):

# Early stop when consistently near max episode length.
if np.mean(scores[-min(10, len(scores)):]) > 490:
torch.save(agent.model.state_dict(), SAVE_PATH)
print(f"Saved trained model to {SAVE_PATH}")
sys.exit()
solved = True
break

torch.save(agent.model.state_dict(), SAVE_PATH)
print(f"Saved trained model to {SAVE_PATH}")
10 changes: 5 additions & 5 deletions 2-cartpole/2-a2c.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -19,7 +19,6 @@
Subtracting V_w(s) is the variance-reduction baseline; using a learned V
(rather than the Monte-Carlo return) is what makes this *actor-critic*.
"""
import sys

import numpy as np
import torch
Expand Down Expand Up @@ -125,8 +124,11 @@ def train_model(self, state, action, reward, next_state, done):
run_test_loop(env, agent.get_action)

scores = []
solved = False

for e in range(EPISODES):
if solved:
break
done = False
score = 0
state, _ = env.reset()
Expand All @@ -149,10 +151,8 @@ def train_model(self, state, action, reward, next_state, done):
scores.append(score)
print(f"episode: {e} score: {score}")
if np.mean(scores[-min(10, len(scores)):]) > 490:
torch.save({"actor": agent.actor.state_dict(),
"critic": agent.critic.state_dict()}, SAVE_PATH)
print(f"Saved trained model to {SAVE_PATH}")
sys.exit()
solved = True
break

torch.save({"actor": agent.actor.state_dict(),
"critic": agent.critic.state_dict()}, SAVE_PATH)
Expand Down
5 changes: 1 addition & 4 deletions 2-cartpole/3-ppo.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -27,7 +27,6 @@

L = L^CLIP - c_v * MSE(V, returns) + c_e * H[pi]
"""
import sys

import numpy as np
import torch
Expand Down Expand Up @@ -211,9 +210,7 @@ def pick(state):
recent = ep_returns[-10:]
print(f"update: {episode} recent_mean_return: {np.mean(recent):.1f} episodes: {len(ep_returns)}")
if len(recent) >= 10 and np.mean(recent) > 490:
torch.save(model.state_dict(), SAVE_PATH)
print(f"Saved trained model to {SAVE_PATH}")
sys.exit()
break

torch.save(model.state_dict(), SAVE_PATH)
print(f"Saved trained model to {SAVE_PATH}")