Add curriculum / hierarchical-RL extension on top of v2#223
Open
AviralDeshratna11 wants to merge 2 commits into
Open
Add curriculum / hierarchical-RL extension on top of v2#223AviralDeshratna11 wants to merge 2 commits into
AviralDeshratna11 wants to merge 2 commits into
Conversation
Adds a practical curriculum-learning system layered on the existing v2 RedGymEnv without modifying it (baseline_fast_v2.py / red_gym_env_v2.py untouched): - curriculum/ram_map.py: single-source RAM reader (GameState) - curriculum/structured_obs.py: normalized structured observation vector + space - curriculum/milestones.py: 33-milestone curriculum graph + MilestoneManager - curriculum/rewards.py: configurable, logged rewards (one-time/max-based) + anti-loop - curriculum/scripted_helpers.py: deterministic dialogue/heal/shop macros - curriculum/skills.py: option-policy scaffold + high-level controller stub - curriculum/state_store.py: per-milestone success-state snapshots (multiproc-safe) - curriculum/config.py: YAML/JSON config dataclasses - curriculum/tb_callback.py: num-envs-agnostic TensorBoard callback - curriculum_env.py: CurriculumRedGymEnv subclass wiring it together - train_full_game_curriculum.py: staged PPO driver (headless + visual modes) - run_curriculum_interactive.py: watch/evaluate a policy - configs/curriculum.yaml, requirements-curriculum.txt, README_curriculum.md No ROM is included or distributed; user provides PokemonRed.gb at the repo root. Co-Authored-By: Claude Opus 4.8 (1M context) <[email protected]>
Owner
|
does it work? |
Author
|
I ran the agent , as of now it was able to defeat Brock and enter the cave
, after that it wasn't able to find the way out so it started to exploit
the battle reward, so in order to stop this I implemented MARL and
increase exploration over exploitation which has given some progress
results,
Will update the model,
Can I reach out to you if I need your advice?
Btw a huge thanks to you for this inspiring work, I want to take you
baseline model and make this rl agent clear the entire game.
…On Sun, 14 Jun, 2026, 23:55 Peter Whidden, ***@***.***> wrote:
*PWhiddy* left a comment (PWhiddy/PokemonRedExperiments#223)
<#223 (comment)>
does it work?
—
Reply to this email directly, view it on GitHub
<#223?email_source=notifications&email_token=ARB74LDRA7WSN5V5PFYIRW3473U3LA5CNFSNUABFM5UWIORPF5TWS5BNNB2WEL2JONZXKZKDN5WW2ZLOOQXTINZQGI3DINRWHA4KM4TFMFZW63VGMF2XI2DPOKSWK5TFNZ2KYZTPN52GK4S7MNWGSY3L#issuecomment-4702646688>,
or unsubscribe
<https://github.com/notifications/unsubscribe-auth/ARB74LCTNS3NGXPXYZR3F4T473U3LAVCNFSNUABFKJSXA33TNF2G64TZHMZDGMBQGY2DOOJWHNEXG43VMU5TINRVGY4DIMRXGQZ2C5QC>
.
Triage notifications, keep track of coding agent tasks and review pull
requests on the go with GitHub Mobile for iOS
<https://github.com/notifications/mobile/ios/ARB74LAXBSB2YZTQUVYKZXL473U3LA5CNFSNUABFM5UWIORPF5TWS5BNNB2WEL2JONZXKZKDN5WW2ZLOOQXTINZQGI3DINRWHA4KM4TFMFZW63VGMF2XI2DPOKSWK5TFNZ2KUZTPN52GK4S7NFXXG>
and Android
<https://github.com/notifications/mobile/android/ARB74LCYYBOELL4WMRVEZ63473U3LA5CNFSNUABFM5UWIORPF5TWS5BNNB2WEL2JONZXKZKDN5WW2ZLOOQXTINZQGI3DINRWHA4KM4TFMFZW63VGMF2XI2DPOKSWK5TFNZ2K4ZTPN52GK4S7MFXGI4TPNFSA>.
Download it today!
You are receiving this because you authored the thread.Message ID:
***@***.***>
|
Owner
|
ofc! If you run v2 with no changes it should eventually get far past that, all the way to ss anne. Feel free to ask questions in the discord! |
Author
|
ohk thanks, the the agent is running, will share the updates when i reach the ss anne and if my pipelines works beyond that |
…when stuck to update the weights and rewards and policy
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Adds a practical curriculum-learning system layered on the existing v2 RedGymEnv without modifying it (baseline_fast_v2.py / red_gym_env_v2.py untouched):
No ROM is included or distributed; user provides PokemonRed.gb at the repo root.