All pull requests

Pull requests

• fix: migrate remaining gym imports to gymnasium (#471)

  #507·EdenRochmanSharabi opened Sep 18, 2026·3 tasks done·

  ·
  1

  1

• WIP: polish(pu): polish unizero+mspacman performance

  #505·puyuan1996 opened Sep 9, 2026Collaborator·

  ·

• fix(model): effective-rank and dormant-ratio logging call sites raise TypeError

  #504·Anai-Guo opened Sep 4, 2026Contributor·

  ·

• fix(gaz): repair Gumbel AlphaZero ctree search

  #502·puyuan1996 opened Aug 25, 2026Collaborator·

  ·
  1

• fix(pu): add batch wrapper for alphazero ptree mcts

  #485·puyuan1996 opened May 31, 2026Collaborator·

  ·

• Fix: rezero buffers variable action space

  #480·SteamMachinist opened Apr 28, 2026Contributor·

  ·
  1

• feature(tj): add MoE expert selection stats and gradient conflict monitoring

  research
  #470·tAnGjIa520 opened Feb 3, 2026Contributor·

  ·
  4

• feature(pu): add use_sigreg option

  research
  #469·puyuan1996 opened Jan 27, 2026Collaborator·

  ·

• feature(xjy): add multi-task learning pipeline in jericho environment

  research
  #465·xiongjyu opened Jan 18, 2026Collaborator·

  ·
  7

• feature(tj): add EfficientZero v2

  algorithm
  #461·tAnGjIa520 opened Jan 5, 2026Contributor·

  ·
  4

• feature(tj): add sequential halving support

  algorithm
  #449·tAnGjIa520 opened Dec 2, 2025Contributor·

  ·

• feature(mycve): add chinese chess env and related demo

  enhancementenvironment
  #442·mycve opened Nov 23, 2025·

  ·
  14

• feature(xjy): add the rnd-related features

  algorithmconfigresearch
  #438·xiongjyu opened Nov 7, 2025Collaborator·

  ·
  9

• feature(tj): add evaluation utils from open-reasoner-zero to priorzero

  research
  #437·tAnGjIa520 opened Nov 7, 2025Contributor·

  ·

• feature(tj): add monitoring for the gradient conflict metric of MoE in ScaleZero

  #421·tAnGjIa520 opened Sep 27, 2025Contributor·

  ·

• feature(xjy): collect data in jericho for rft

  enhancementresearch
  #420·xiongjyu opened Sep 26, 2025Collaborator·

  ·

• fix(xjy): adding the messenger environment

  environmentresearch
  #405·xiongjyu opened Aug 18, 2025Collaborator·

  ·
  4

• fix(tj): finetune spaceinvaders from atari26 pretrained ckpt in ScaleZero

  enhancementresearch
  #399·tAnGjIa520 opened Aug 12, 2025Contributor·

  ·
  6

• WIP: polish(pu): add a polished version of qwen prior policy

  #397·puyuan1996 opened Aug 11, 2025Collaborator·

  ·

• WIP: feature(nyz/pu): add init version of async demo using task pipeline

  #396·puyuan1996 opened Aug 5, 2025Collaborator·

  ·

• WIP: feature(pu): add init version of async unizero using multi-threading

  #395·puyuan1996 opened Aug 1, 2025Collaborator·

  ·

• fix(pu): fix chess reset bug when use alphazero ctree

  #364·puyuan1996 opened May 23, 2025Collaborator·

  ·

• How to fix the bug of loading trained model for evaluation

  #340·xiongjyu opened Apr 2, 2025Collaborator·

  ·

• feature(xjy): add mamba2 as a unizero backbone option

  algorithm
  #338·xiongjyu opened Mar 31, 2025Collaborator·

  ·
  1

• WIP: feature(pu): add muzero with history encoder

  algorithmenhancement
  #334·puyuan1996 opened Mar 21, 2025Collaborator·

  ·