Merge branch 'main' into feature/embodied_unify_interface

Author: andylin-hao

README.md

@@ -30,6 +30,8 @@ RLinf is a flexible and scalable open-source infrastructure designed for post-tr
 
 
 ## What's NEW!
+- [2025/11] 🔥 RLinf supports reinforcement learning fine-tuning for [IsaacLab](https://github.com/isaac-sim/IsaacLab). Doc: [RL on IsaacLab](https://rlinf.readthedocs.io/en/latest/rst_source/examples/isaaclab.html) 
+- [2025/11] 🔥 RLinf supports reinforcement learning fine-tuning for [Behavior 1k](https://github.com/StanfordVL/BEHAVIOR-1K). Doc: [RL on Behavior 1k](https://rlinf.readthedocs.io/en/latest/rst_source/examples/behavior.html) 
 - [2025/11] 🔥 RLinf supports reinforcement learning fine-tuning for [GR00T-N1.5](https://github.com/NVIDIA/Isaac-GR00T). Doc: [RL on GR00T-N1.5](https://rlinf.readthedocs.io/en/latest/rst_source/examples/gr00t.html).
 - [2025/11] 🔥 RLinf supports reinforcement learning fine-tuning for [Metaworld](https://github.com/Farama-Foundation/Metaworld). Doc: [RL on Metaworld](https://rlinf.readthedocs.io/en/latest/rst_source/examples/metaworld.html).
 - [2025/11] 🔥 RLinf supports reinforcement learning fine-tuning for [Behavior 1k](https://github.com/StanfordVL/BEHAVIOR-1K). Doc: [RL on Behavior 1k](https://rlinf.readthedocs.io/en/latest/rst_source/examples/behavior.html).
@@ -66,7 +68,7 @@ RLinf is a flexible and scalable open-source infrastructure designed for post-tr
           <li>RoboVerse</li>
           <li><a href="https://rlinf.readthedocs.io/en/latest/rst_source/examples/behavior.html">BEHAVIOR</a> ✅</li>
           <li><a href="https://rlinf.readthedocs.io/en/latest/rst_source/examples/metaworld.html">MetaWorld</a> ✅</li>
-          <li>IsaacLab</li>
+          <li><a href="https://rlinf.readthedocs.io/en/latest/rst_source/examples/isaaclab.html">IsaacLab</a> ✅</li>
           <li>RoboCasa</li>
           <li>More...</li>
         </ul>
@@ -199,7 +201,7 @@ and exhibits greater stability.
     <td style="text-align:center;">39.10%</td>
   </tr>
   <tr>
-    <td style="text-align:center;"><a href="https://huggingface.co/gen-robot/openvla-7b-rlvla-warmup"><img src="docs/source-en/_static/svg/hf-logo.svg" alt="HF" width="16" height="16" style="vertical-align: middle;">RL4VLA (PPO)</a></td>
+    <td style="text-align:center;"><a href="https://huggingface.co/gen-robot/openvla-7b-rlvla-rl"><img src="docs/source-en/_static/svg/hf-logo.svg" alt="HF" width="16" height="16" style="vertical-align: middle;">RL4VLA (PPO)</a></td>
     <td style="text-align:center;">93.75%</td>
     <td style="text-align:center;">80.47%</td>
     <td style="text-align:center;">75.00%</td>

README.zh-CN.md

@@ -30,6 +30,8 @@ RLinf 是一个灵活且可扩展的开源框架，专为利用强化学习进
 
 
 ## 最新动态
+- [2025/11] 🔥 基于[IsaacLab](https://github.com/isaac-sim/IsaacLab)的强化学习微调已经上线! 文档：[RL on IsaacLab](https://rlinf.readthedocs.io/zh-cn/latest/rst_source/examples/isaaclab.html) 
+- [2025/11] 🔥 基于[Behavior 1k](https://github.com/StanfordVL/BEHAVIOR-1K)的强化学习微调已经上线! 文档：[RL on Behavior 1k](https://rlinf.readthedocs.io/zh-cn/latest/rst_source/examples/behavior.html) 
 - [2025/11] 🔥 RLinf现在已经支持强化学习微调[GR00T-N1.5](https://github.com/NVIDIA/Isaac-GR00T)！文档：[RL on GR00T-N1.5](https://rlinf.readthedocs.io/zh-cn/latest/rst_source/examples/gr00t.html)。
 - [2025/11] 🔥 基于[Metaworld](https://github.com/Farama-Foundation/Metaworld)的强化学习微调已经上线! 文档：[RL on Metaworld](https://rlinf.readthedocs.io/zh-cn/latest/rst_source/examples/metaworld.html)。
 - [2025/11] 🔥 基于[Behavior 1k](https://github.com/StanfordVL/BEHAVIOR-1K)的强化学习微调已经上线! 文档：[RL on Behavior 1k](https://rlinf.readthedocs.io/zh-cn/latest/rst_source/examples/behavior.html) 。
@@ -66,7 +68,7 @@ RLinf 是一个灵活且可扩展的开源框架，专为利用强化学习进
           <li>RoboVerse</li>
           <li><a href="https://rlinf.readthedocs.io/zh-cn/latest/rst_source/examples/behavior.html">BEHAVIOR</a> ✅</li>
           <li><a href="https://rlinf.readthedocs.io/zh-cn/latest/rst_source/examples/metaworld.html">MetaWorld</a> ✅</li>
-          <li>IsaacLab</li>
+          <li><a href="https://rlinf.readthedocs.io/zh-cn/latest/rst_source/examples/isaaclab.html">IsaacLab</a> ✅</li>
           <li>RoboCasa</li>
           <li>More...</li>
         </ul>
@@ -197,7 +199,7 @@ RLinf 是一个灵活且可扩展的开源框架，专为利用强化学习进
     <td style="text-align:center;">39.10%</td>
   </tr>
   <tr>
-    <td style="text-align:center;"><a href="https://huggingface.co/gen-robot/openvla-7b-rlvla-warmup"><img src="docs/source-en/_static/svg/hf-logo.svg" alt="HF" width="16" height="16" style="vertical-align: middle;">RL4VLA (PPO)</a></td>
+    <td style="text-align:center;"><a href="https://huggingface.co/gen-robot/openvla-7b-rlvla-rl"><img src="docs/source-en/_static/svg/hf-logo.svg" alt="HF" width="16" height="16" style="vertical-align: middle;">RL4VLA (PPO)</a></td>
     <td style="text-align:center;">93.75%</td>
     <td style="text-align:center;">80.47%</td>
     <td style="text-align:center;">75.00%</td>

docs/source-en/index.rst

@@ -40,7 +40,7 @@ RLinf is a flexible and scalable open-source infrastructure designed for post-tr
 - Embodied Agent Support
 
   - Fast adaptation support for mainstream VLA models: `OpenVLA`_, `OpenVLA-OFT`_, `π₀`_, `GR00T-N1.5`_
-  - Support for mainstream CPU & GPU-based simulators via standardized RL interfaces: `ManiSkill3`_, `LIBERO`_
+  - Support for mainstream CPU & GPU-based simulators via standardized RL interfaces: `ManiSkill3`_, `LIBERO`_, `IsaacLab`_
   - Enabling the first RL fine-tuning of the π₀ model family with a flow-matching action expert.
 
 **RLinf is fast with:**
@@ -73,6 +73,7 @@ RLinf is a flexible and scalable open-source infrastructure designed for post-tr
 .. _IsaacLab: https://github.com/isaac-sim/IsaacLab
 .. _ManiSkill3: https://github.com/haosulab/ManiSkill
 .. _LIBERO: https://github.com/Lifelong-Robot-Learning/LIBERO
+.. _IsaacLab: https://github.com/isaac-sim/IsaacLab
 .. _π₀: https://github.com/Physical-Intelligence/openpi
 .. _Megatron-LM: https://github.com/NVIDIA/Megatron-LM
 .. _SGLang: https://github.com/sgl-project/sglang

docs/source-en/rst_source/examples/index.rst

@@ -74,6 +74,17 @@ as well as reinforcement learning training examples on real robots.
        </p>
      </div>
 
+      <div style="flex: 1 1 30%; max-width: 300px; text-align: center;">
+       <img src="https://github.com/RLinf/misc/raw/main/pic/IsaacLab.png"
+            style="width: 100%; height: 200px; object-fit: cover; border-radius: 8px; box-shadow: 0 2px 6px rgba(0,0,0,0.15);" />
+       <p style="margin-top: 8px; font-size: 14px; line-height: 1.4;">
+         <a href="https://rlinf.readthedocs.io/en/latest/rst_source/examples/isaaclab.html" target="_blank" style="text-decoration: underline; color: blue;">
+          <b>RL with IsaacLab Benchmark</b>
+         </a><br>
+         Support IsaacLab+gr00t+PPO training
+       </p>
+     </div>
+  
       <div style="flex: 1 1 30%; max-width: 300px; text-align: center;">
        <img src="https://github.com/RLinf/misc/raw/main/pic/gr00t.png"
             style="width: 100%; height: 200px; object-fit: cover; border-radius: 8px; box-shadow: 0 2px 6px rgba(0,0,0,0.15);" />
@@ -235,6 +246,7 @@ Thanks to this decoupled design, workers can be flexibly and dynamically schedul
    libero
    behavior
    metaworld
+   isaaclab
    pi0
    gr00t
    reasoning