Skip to content

Promote an MMD-trained v2 policy to the browser champion #1

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

Open
iakraptis wants to merge 1 commit into
base: main
Choose a base branch
from
Open

Promote an MMD-trained v2 policy to the browser champion #1

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
13 changes: 8 additions & 5 deletions policy/README.md
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -1,14 +1,17 @@
# Browser policy

`static-policy.json` is the browser-native actor exported from the validated
`bang_structured_v2` checkpoint at iteration 2995. Keeping the inference
artifact in the repository makes the production build reproducible in GitHub
Actions without granting CI access to local training runs.
`static-policy.json` is the browser-native actor exported from a
`bang_structured_v2` checkpoint at iteration 2861 (run `mmd_v2_a030_s2`). It was
trained with the standard v2 League PPO recipe plus a Magnetic Mirror Descent
term (alpha 0.3, magnet moved every 50 updates), and replaces the iteration-2995
PPO champion. Keeping the inference artifact in the repository makes the
production build reproducible in GitHub Actions without granting CI access to
local training runs.

Regenerate it from a complete checkpoint with:

```sh
python3 tools/export_static_policy.py \
runs/bang_league_ppo/main/latest.pt \
runs/mmd_v2_a030_s2/champion/policy.pt \
policy/static-policy.json
```
2 changes: 1 addition & 1 deletion policy/static-policy.json
View file
Open in desktop

Large diffs are not rendered by default.