Merge branch 'opendilab:main' into dev-unizero-mamba2

Author: xiongjyu

.github/workflows/release.yml

@@ -38,10 +38,10 @@ jobs:
         run: |
           make zip
           ls -al dist
-      - name: Upload packed files to artifacts
-        uses: actions/upload-artifact@v2
+      - name: Upload packed files to artifacts (source)
+        uses: actions/upload-artifact@v4
         with:
-          name: build-artifacts-all
+          name: build-artifacts-source
           path: ./dist/*
 
   wheel_build:
@@ -102,35 +102,119 @@ jobs:
         run: |
           ls -al ./wheelhouse
           mv wheelhouse dist
-      - name: Upload packed files to artifacts
-        uses: actions/upload-artifact@v2
+      - name: Upload packed files to artifacts (wheels)
+        uses: actions/upload-artifact@v4
         with:
-          name: build-artifacts-all
+          name: build-artifacts-wheels-${{ matrix.os }}-${{ matrix.python }}-${{ matrix.architecture }}
           path: ./dist/*
 
+  wheel_aggregate:
+    name: Aggregate all wheels
+    runs-on: ubuntu-20.04
+    needs: wheel_build
+    steps:
+      - name: Create aggregation directory
+        run: mkdir -p aggregated_wheels_all
+
+      - name: Download wheel ubuntu-20.04, 3.7, x86_64
+        uses: actions/download-artifact@v4
+        with:
+          name: build-artifacts-wheels-ubuntu-20.04-3.7-x86_64
+          path: aggregated_wheels_all
+      - name: Download wheel ubuntu-20.04, 3.8, x86_64
+        uses: actions/download-artifact@v4
+        with:
+          name: build-artifacts-wheels-ubuntu-20.04-3.8-x86_64
+          path: aggregated_wheels_all
+      - name: Download wheel ubuntu-20.04, 3.9, x86_64
+        uses: actions/download-artifact@v4
+        with:
+          name: build-artifacts-wheels-ubuntu-20.04-3.9-x86_64
+          path: aggregated_wheels_all
+      - name: Download wheel ubuntu-20.04, 3.10, x86_64
+        uses: actions/download-artifact@v4
+        with:
+          name: build-artifacts-wheels-ubuntu-20.04-3.10-x86_64
+          path: aggregated_wheels_all
+      - name: Download wheel ubuntu-20.04, 3.11, x86_64
+        uses: actions/download-artifact@v4
+        with:
+          name: build-artifacts-wheels-ubuntu-20.04-3.11-x86_64
+          path: aggregated_wheels_all
+      - name: Download wheel ubuntu-20.04, 3.7, aarch64
+        uses: actions/download-artifact@v4
+        with:
+          name: build-artifacts-wheels-ubuntu-20.04-3.7-aarch64
+          path: aggregated_wheels_all
+      - name: Download wheel ubuntu-20.04, 3.8, aarch64
+        uses: actions/download-artifact@v4
+        with:
+          name: build-artifacts-wheels-ubuntu-20.04-3.8-aarch64
+          path: aggregated_wheels_all
+      - name: Download wheel ubuntu-20.04, 3.9, aarch64
+        uses: actions/download-artifact@v4
+        with:
+          name: build-artifacts-wheels-ubuntu-20.04-3.9-aarch64
+          path: aggregated_wheels_all
+      - name: Download wheel ubuntu-20.04, 3.10, aarch64
+        uses: actions/download-artifact@v4
+        with:
+          name: build-artifacts-wheels-ubuntu-20.04-3.10-aarch64
+          path: aggregated_wheels_all
+      - name: Download wheel ubuntu-20.04, 3.11, aarch64
+        uses: actions/download-artifact@v4
+        with:
+          name: build-artifacts-wheels-ubuntu-20.04-3.11-aarch64
+          path: aggregated_wheels_all
+
+      - name: Download wheel macos-13, 3.7, x86_64
+        uses: actions/download-artifact@v4
+        with:
+          name: build-artifacts-wheels-macos-13-3.7-x86_64
+          path: aggregated_wheels_all
+      - name: Download wheel macos-13, 3.8, x86_64
+        uses: actions/download-artifact@v4
+        with:
+          name: build-artifacts-wheels-macos-13-3.8-x86_64
+          path: aggregated_wheels_all
+      - name: Download wheel macos-13, 3.7, arm64
+        uses: actions/download-artifact@v4
+        with:
+          name: build-artifacts-wheels-macos-13-3.7-arm64
+          path: aggregated_wheels_all
+      - name: Download wheel macos-13, 3.8, arm64
+        uses: actions/download-artifact@v4
+        with:
+          name: build-artifacts-wheels-macos-13-3.8-arm64
+          path: aggregated_wheels_all
+
+      - name: Upload unified wheels artifact
+        uses: actions/upload-artifact@v4
+        with:
+          name: build-artifacts-wheels
+          path: aggregated_wheels_all
+
   # the publishing can only be processed on linux system
   wheel_publish:
     name: Publish the wheels to pypi
     runs-on: ubuntu-20.04
     needs:
       - source_build
-      - wheel_build
+      - wheel_aggregate
     strategy:
       fail-fast: false
       matrix:
         python:
           - '3.8.7'
 
     steps:
-      - name: Download packed files to artifacts
-        uses: actions/download-artifact@v3
+      - name: Download unified wheels artifact
+        uses: actions/download-artifact@v4
         with:
-          name: build-artifacts-all
+          name: build-artifacts-wheels
           path: ./dist
-      - name: Show the buildings
-        shell: bash
-        run: |
-          ls -al ./dist
+      - name: Show the aggregated wheels
+        run: ls -al ./dist
       - name: Upload distribution 📦 to github release
         uses: svenstaro/upload-release-action@v2
         with:
@@ -145,4 +229,4 @@ jobs:
           password: ${{ secrets.PYPI_API_TOKEN }}
           verbose: true
           skip_existing: true
-          packages_dir: dist/
+          packages_dir: dist/

.github/workflows/release_test.yml

@@ -38,12 +38,12 @@ jobs:
           make zip
           ls -al dist
       - name: Upload packed files to artifacts
-        uses: actions/upload-artifact@v2
+        uses: actions/upload-artifact@v4
         with:
           name: build-artifacts-source-pack
           path: ./dist/*
       - name: Upload packed files to artifacts
-        uses: actions/upload-artifact@v2
+        uses: actions/upload-artifact@v4
         with:
           name: build-artifacts-all
           path: ./dist/*
@@ -108,12 +108,12 @@ jobs:
           ls -al ./wheelhouse
           mv wheelhouse dist
       - name: Upload packed files to artifacts
-        uses: actions/upload-artifact@v3
+        uses: actions/upload-artifact@v4
         with:
           name: build-artifacts-${{ runner.os }}-cp${{ matrix.python }}-${{ matrix.architecture }}
           path: ./dist/*
       - name: Upload packed files to artifacts
-        uses: actions/upload-artifact@v2
+        uses: actions/upload-artifact@v4
         with:
           name: build-artifacts-all
           path: ./dist/*

.github/workflows/test.yml

@@ -100,6 +100,7 @@ jobs:
         run: |
             make clean build unittest
       - name: Upload coverage to Codecov
+        if: ${{ env.OS_NAME == 'MacOS' }}
         uses: codecov/codecov-action@v4
         with:
           token: ${{ secrets.CODECOV_TOKEN }}

.gitignore

@@ -1449,3 +1449,5 @@ events.*
 # pooltool-specific stuff
 !/assets/pooltool/**
 lzero/mcts/ctree/ctree_alphazero/pybind11
+
+zoo/jericho/envs/z-machine-games-master

CHANGELOG.md

@@ -1,3 +1,41 @@
+2025.04.09 (v0.2.0)
+- env: Add Metadrive environment and configurations (#192)
+- env: Add Sampled MuZero/UniZero and DMC environment with related configurations (#260)
+- env: Polish Chess environment and its render method; add unittests and configurations (#272)
+- env: Add Jericho environment and its configurations (#307)
+- algo: Add Harmony Dream loss balance in MuZero (#242)
+- algo: Adopt AlphaZero for non-zero-sum games (#245)
+- algo: Add AlphaZero CTree unittest (#306)
+- algo: Add recent MCTS-related papers (#324)
+- algo: Introduce rope to use true timestep index as pos_index (#266)
+- algo: Add Jericho DDP configuration (#337)
+- feat: Add LightZero Sphinx documentation (#237)
+- feat: Add Wandb support (#294)
+- feat: Add Atari100k metric utilities (#295)
+- feat: Add eval_benchmark tests (#296)
+- feat: Add save_replay and collect_episode_data options in Jericho (#333)
+- feat: Add an MCTS TicTacToe demo in one single file (#315)
+- fix: Fix DownSample for different observation shapes (#254)
+- fix: Fix wrong chance values in Stochastic MuZero (#275)
+- fix: Use display_frames_as_gif in CartPole (#288)
+- fix: Fix chance encoder in stochastic_muzero_model_mlp.py (#284)
+- fix: Correct typo in model/utils.py (#290)
+- fix: Fix SMZ compile_args and num_simulations bug in world_model (#297)
+- fix: Fix reward type bug in 2048 and OS import issue in CartPole (#304)
+- fix: Switch to macos-13 in action (#319)
+- fix: Fix SMZ & SEZ config for pixel-based DMC (#322)
+- fix: Fix update_per_collect in DDP setting (#321)
+- fix: Fix obs_shape tuple bug in initialize_zeros_batch (#327)
+- fix: Fix prepare_obs_stack_for_unizero (#328)
+- fix: Fix random_policy when len(ready_env_id)<collector_env_num (#335)
+- fix: Fix timestep compatibility (#339)
+- polish: Polish efficiency and performance on Atari and DMC (#292)
+- polish: Update requirements (#298)
+- polish: Optimize reward/value/policy_head_hidden_channels (#314)
+- polish: Update tutorial configuration and log instructions (#330)
+- ci: Add self-hosted Linux (Ubuntu) CI runner (#259)
+- test: Add self-hosted Linux runner for CI tests (#323)
+  
 2024.07.12 (v0.1.0)
 - env: SumToThree env from pooltool(#227)
 - algo: UniZero (#232)

README.md

@@ -28,7 +28,7 @@
 [![GitHub license](https://img.shields.io/github/license/opendilab/LightZero)](https://github.com/opendilab/LightZero/blob/master/LICENSE)
 [![discord badge](https://dcbadge.vercel.app/api/server/dkZS2JF56X?style=flat)](https://discord.gg/dkZS2JF56X)
 
-Updated on 2025.02.08 LightZero-v0.1.0
+Updated on 2025.06.03 LightZero-v0.2.0
 
 English | [简体中文(Simplified Chinese)](https://github.com/opendilab/LightZero/blob/main/README.zh.md) | [Documentation](https://opendilab.github.io/LightZero) | [LightZero Paper](https://arxiv.org/abs/2310.08348) | [🔥UniZero Paper](https://arxiv.org/abs/2406.10667) | [🔥ReZero Paper](https://arxiv.org/abs/2404.16364)
 

README.zh.md

@@ -27,7 +27,7 @@
 [![Contributors](https://img.shields.io/github/contributors/opendilab/LightZero)](https://github.com/opendilab/LightZero/graphs/contributors)
 [![GitHub license](https://img.shields.io/github/license/opendilab/LightZero)](https://github.com/opendilab/LightZero/blob/master/LICENSE)
 
-最近更新于 2025.02.08 LightZero-v0.1.0
+最近更新于 2025.06.03 LightZero-v0.2.0
 
 [English](https://github.com/opendilab/LightZero/blob/main/README.md) | 简体中文 | [文档](https://opendilab.github.io/LightZero) | [LightZero 论文](https://arxiv.org/abs/2310.08348) | [🔥UniZero 论文](https://arxiv.org/abs/2406.10667) | [🔥ReZero 论文](https://arxiv.org/abs/2404.16364)
 
@@ -52,25 +52,37 @@
 **LightZero** 的目标是**标准化 MCTS 算法族，以加速相关研究和应用。** [Benchmark](#benchmark) 中介绍了目前所有已实现算法的性能比较。
 
 ### 导航
-- [概览](#概览)
+- [LightZero](#lightzero)
+  - [🔍 背景](#-背景)
+  - [🎨 概览](#-概览)
     - [导航](#导航)
-    - [特点](#特点)
-    - [框架结构](#框架结构)
-    - [集成算法](#集成算法)
-- [安装方法](#安装方法)
-- [快速开始](#快速开始)
-- [文档](#文档)
-- [基线算法比较](#基线算法比较)
-- [MCTS相关笔记](#MCTS-相关笔记)
+    - [💥 特点](#-特点)
+    - [🧩 框架结构](#-框架结构)
+    - [🎁 集成算法](#-集成算法)
+  - [⚙️ 安装方法](#️-安装方法)
+    - [使用 Docker 进行安装](#使用-docker-进行安装)
+  - [🚀 快速开始](#-快速开始)
+  - [📚 文档](#-文档)
+  - [📊 基线算法比较](#-基线算法比较)
+  - [📝 MCTS 相关笔记](#-mcts-相关笔记)
     - [论文笔记](#论文笔记)
     - [算法框架图](#算法框架图)
-- [MCTS相关论文](#MCTS-相关论文)
+  - [MCTS 相关论文](#mcts-相关论文)
     - [重要论文](#重要论文)
+      - [LightZero Implemented series](#lightzero-implemented-series)
+      - [AlphaGo series](#alphago-series)
+      - [MuZero series](#muzero-series)
+      - [MCTS Analysis](#mcts-analysis)
+      - [MCTS Application](#mcts-application)
     - [其他论文](#其他论文)
-- [反馈意见和贡献](#反馈意见和贡献)
-- [引用](#引用)
-- [致谢](#致谢)
-- [许可证](#许可证)
+      - [ICML](#icml)
+      - [ICLR](#iclr)
+      - [NeurIPS](#neurips)
+      - [Other Conference or Journal](#other-conference-or-journal)
+  - [💬 反馈意见和贡献](#-反馈意见和贡献)
+  - [🌏 引用](#-引用)
+  - [💓 致谢](#-致谢)
+  - [🏷️ 许可证](#️-许可证)
 
 ### 💥 特点
 **轻量**：LightZero 中集成了多种 MCTS 族算法，能够在同一框架下轻量化地解决多种属性的决策问题。

docs/source/tutorials/envs/customize_envs.md

@@ -81,12 +81,17 @@ In a custom environment, you need to provide properties for observation space an
 
 ```python
 @property
-defobservation_space(self):
-    return self.env.observation_space
+def observation_space(self):
+    return self._observation_space
 
 @property
 def action_space(self):
-    return self.env.action_space
+    return self._action_space
+    
+@property
+def legal_actions(self):
+    # get the actual legal actions
+    return np.arange(self._action_space.n)
 ```
 
 ### 6. **Render Method**<br>

lzero/config/meta.py

@@ -7,7 +7,7 @@
 __TITLE__ = "LightZero"
 
 #: Version of this project.
-__VERSION__ = "0.1.0"
+__VERSION__ = "0.2.0"
 
 #: Short description of the project, will be included in ``setup.py``.
 __DESCRIPTION__ = 'A lightweight and efficient MCTS/AlphaZero/MuZero algorithm toolkits.'

lzero/entry/__init__.py

@@ -1,5 +1,6 @@
 from .eval_alphazero import eval_alphazero
 from .eval_muzero import eval_muzero
+
 from .eval_muzero_with_gym_env import eval_muzero_with_gym_env
 from .train_alphazero import train_alphazero
 from .train_muzero import train_muzero