rlinf 0.2.0.dev1__tar.gz → 0.2.0.dev2__tar.gz

{rlinf-0.2.0.dev1/rlinf.egg-info → rlinf-0.2.0.dev2}/PKG-INFO

@@ -1,6 +1,8 @@
 Metadata-Version: 2.4
 Name: rlinf
-Version: 0.2.0.dev1
+Version: 0.2.0.dev2
+Summary: Reinforcement Learning Infrastructure for Embodied and Agentic AI
+Author: RLinf Contributors
 License:                                  Apache License
                                    Version 2.0, January 2004
                                 https://www.apache.org/licenses/
@@ -202,6 +204,10 @@ License:                                  Apache License
            WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
            See the License for the specific language governing permissions and
            limitations under the License.
+Project-URL: Homepage, https://github.com/RLinf/RLinf
+Project-URL: Repository, https://github.com/RLinf/RLinf
+Project-URL: Documentation, https://rlinf.readthedocs.io/
+Project-URL: Bug Report, https://github.com/RLinf/RLinf/issues
 Keywords: reinforcement-learning,embodied-intelligence,large-language-models
 Classifier: Development Status :: 2 - Pre-Alpha
 Classifier: Environment :: GPU :: NVIDIA CUDA :: 12 :: 12.4
@@ -212,10 +218,10 @@ Requires-Python: <=3.11.14,>=3.10
 Description-Content-Type: text/markdown
 License-File: LICENSE
 Requires-Dist: ray[default]>=2.47.0
-Requires-Dist: torch==2.6.0
+Requires-Dist: torch<=2.9.0,>=2.5.0
 Requires-Dist: hydra-core
 Requires-Dist: numpy
-Requires-Dist: datasets==3.6.0
+Requires-Dist: datasets
 Requires-Dist: torchdata
 Requires-Dist: scipy
 Requires-Dist: accelerate
@@ -223,7 +229,7 @@ Requires-Dist: debugpy
 Requires-Dist: einops
 Requires-Dist: nvitop
 Requires-Dist: pybind11
-Requires-Dist: setuptools<75.9,>=69.5.1
+Requires-Dist: setuptools
 Requires-Dist: ninja
 Requires-Dist: pytest
 Requires-Dist: gsutil
@@ -241,44 +247,35 @@ Requires-Dist: tensorboard
 Provides-Extra: agentic-sglang
 Requires-Dist: sglang[all]==0.4.6.post5; extra == "agentic-sglang"
 Requires-Dist: torch-memory-saver; extra == "agentic-sglang"
+Requires-Dist: numpy==2.2; extra == "agentic-sglang"
 Requires-Dist: transformers==4.51.1; extra == "agentic-sglang"
 Requires-Dist: uvloop==0.21.0; extra == "agentic-sglang"
 Requires-Dist: hydra-core==1.4.0.dev1; extra == "agentic-sglang"
+Requires-Dist: omegaconf==2.4.0.dev4; extra == "agentic-sglang"
 Requires-Dist: word2number; extra == "agentic-sglang"
 Requires-Dist: regex; extra == "agentic-sglang"
 Requires-Dist: peft==0.11.1; extra == "agentic-sglang"
 Requires-Dist: rlinf_latex2sympy2; extra == "agentic-sglang"
+Requires-Dist: mcp; extra == "agentic-sglang"
 Provides-Extra: agentic-vllm
 Requires-Dist: vllm==0.8.5; extra == "agentic-vllm"
+Requires-Dist: numpy==2.2; extra == "agentic-vllm"
 Requires-Dist: transformers==4.51.1; extra == "agentic-vllm"
 Requires-Dist: uvloop==0.21.0; extra == "agentic-vllm"
 Requires-Dist: hydra-core==1.4.0.dev1; extra == "agentic-vllm"
+Requires-Dist: omegaconf==2.4.0.dev4; extra == "agentic-vllm"
 Requires-Dist: word2number; extra == "agentic-vllm"
 Requires-Dist: regex; extra == "agentic-vllm"
 Requires-Dist: peft==0.11.1; extra == "agentic-vllm"
 Requires-Dist: rlinf_latex2sympy2; extra == "agentic-vllm"
+Requires-Dist: mcp; extra == "agentic-vllm"
 Provides-Extra: embodied
-Requires-Dist: transformers==4.40.1; extra == "embodied"
-Requires-Dist: draccus; extra == "embodied"
-Requires-Dist: rich; extra == "embodied"
-Requires-Dist: tensorflow; extra == "embodied"
-Requires-Dist: tensorflow_graphics; extra == "embodied"
-Requires-Dist: peft==0.11.1; extra == "embodied"
-Requires-Dist: timm==0.9.10; extra == "embodied"
-Requires-Dist: sapien==3.0.1; platform_system == "Linux" and extra == "embodied"
-Requires-Dist: tensordict; extra == "embodied"
+Requires-Dist: transformers<=4.57.6; extra == "embodied"
+Requires-Dist: peft; extra == "embodied"
+Requires-Dist: timm; extra == "embodied"
 Requires-Dist: imageio[ffmpeg]; extra == "embodied"
-Requires-Dist: robosuite==1.4.1; extra == "embodied"
-Requires-Dist: bddl; extra == "embodied"
-Requires-Dist: easydict; extra == "embodied"
-Requires-Dist: cloudpickle; extra == "embodied"
-Requires-Dist: gym; extra == "embodied"
 Requires-Dist: gymnasium; extra == "embodied"
-Requires-Dist: av; extra == "embodied"
-Requires-Dist: open3d; extra == "embodied"
-Requires-Dist: zarr; extra == "embodied"
-Requires-Dist: openai; extra == "embodied"
-Requires-Dist: jax; extra == "embodied"
+Requires-Dist: gym; extra == "embodied"
 Provides-Extra: franka
 Requires-Dist: pyyaml; extra == "franka"
 Requires-Dist: rospkg; extra == "franka"
@@ -294,7 +291,7 @@ Requires-Dist: pynput; extra == "franka"
 Dynamic: license-file
 
 <div align="center">
-  <img src="docs/source-en/_static/svg/logo_white.svg" alt="RLinf-logo" width="600"/>
+  <img src="https://github.com/RLinf/misc/raw/main/pic/logo_white.svg" alt="RLinf-logo" width="600"/>
 </div>
 
 <div align="center">
@@ -320,7 +317,7 @@ Dynamic: license-file
 RLinf is a flexible and scalable open-source RL infrastructure designed for Embodied and Agentic AI. The 'inf' in RLinf stands for `Infrastructure`, highlighting its role as a robust backbone for next-generation training. It also stands for `Infinite`, symbolizing the system’s support for open-ended learning, continuous generalization, and limitless possibilities in intelligence development.
 
 <div align="center">
-  <img src="docs/source-en/_static/svg/overview.svg" alt="RLinf-overview"/>
+  <img src="https://github.com/RLinf/misc/raw/main/pic/overview.svg" alt="RLinf-overview"/>
 </div>
 
 
@@ -417,7 +414,7 @@ Multiple Backend Integrations
           </ul>
           <li><b>Custom Models</b></li>
           <ul>
-            <li><a href="https://github.com/RLinf/RLinf/blob/main/docs/source-en/rst_source/examples/embodied/mlp.rst">MLP-Policy</a> ✅</li>
+            <li><a href="https://rlinf.readthedocs.io/en/latest/rst_source/examples/embodied/mlp.html">MLP-Policy</a> ✅</li>
             <li>CNN-Policy ✅</li>
           </ul>
         </ul>

{rlinf-0.2.0.dev1 → rlinf-0.2.0.dev2}/README.md

@@ -1,5 +1,5 @@
 <div align="center">
-  <img src="docs/source-en/_static/svg/logo_white.svg" alt="RLinf-logo" width="600"/>
+  <img src="https://github.com/RLinf/misc/raw/main/pic/logo_white.svg" alt="RLinf-logo" width="600"/>
 </div>
 
 <div align="center">
@@ -25,7 +25,7 @@
 RLinf is a flexible and scalable open-source RL infrastructure designed for Embodied and Agentic AI. The 'inf' in RLinf stands for `Infrastructure`, highlighting its role as a robust backbone for next-generation training. It also stands for `Infinite`, symbolizing the system’s support for open-ended learning, continuous generalization, and limitless possibilities in intelligence development.
 
 <div align="center">
-  <img src="docs/source-en/_static/svg/overview.svg" alt="RLinf-overview"/>
+  <img src="https://github.com/RLinf/misc/raw/main/pic/overview.svg" alt="RLinf-overview"/>
 </div>
 
 
@@ -122,7 +122,7 @@ Multiple Backend Integrations
           </ul>
           <li><b>Custom Models</b></li>
           <ul>
-            <li><a href="https://github.com/RLinf/RLinf/blob/main/docs/source-en/rst_source/examples/embodied/mlp.rst">MLP-Policy</a> ✅</li>
+            <li><a href="https://rlinf.readthedocs.io/en/latest/rst_source/examples/embodied/mlp.html">MLP-Policy</a> ✅</li>
             <li>CNN-Policy ✅</li>
           </ul>
         </ul>

{rlinf-0.2.0.dev1 → rlinf-0.2.0.dev2}/pyproject.toml

@@ -4,7 +4,9 @@ build-backend = "setuptools.build_meta"
 
 [project]
 name = "rlinf"
-version = "0.2.0.dev1"
+version = "0.2.0.dev2"
+authors = [{ name = "RLinf Contributors" }]
+description = "Reinforcement Learning Infrastructure for Embodied and Agentic AI"
 readme = {file = "README.md", content-type = "text/markdown"}
 requires-python = ">=3.10,<=3.11.14"
 license = {file = "LICENSE"}
@@ -25,16 +27,17 @@ classifiers = [
     "Programming Language :: Python :: 3.11",
 ]
 
+
 dependencies = [
     # Core System
     # These are the dependencies of the core scheduler
     "ray[default]>=2.47.0",
-    "torch==2.6.0",
+    "torch>=2.5.0,<=2.9.0",
     "hydra-core",
     "numpy",
 
     # Data processing
-    "datasets==3.6.0",
+    "datasets",
     "torchdata",
     "scipy",
 
@@ -44,7 +47,7 @@ dependencies = [
     "einops",
     "nvitop",
     "pybind11",
-    "setuptools>=69.5.1,<75.9",
+    "setuptools",
     "ninja",
     "pytest",
     "gsutil",
@@ -63,52 +66,49 @@ dependencies = [
     "tensorboard",
 ]
 
+[project.urls]
+Homepage = "https://github.com/RLinf/RLinf"
+Repository = "https://github.com/RLinf/RLinf"
+Documentation = "https://rlinf.readthedocs.io/"
+"Bug Report" = "https://github.com/RLinf/RLinf/issues"
+
 [project.optional-dependencies]
 agentic-sglang = [
     "sglang[all]==0.4.6.post5",
     "torch-memory-saver",
+    "numpy==2.2",
     "transformers==4.51.1",
     # uvloop is not compatible with simulators, put it here
     "uvloop==0.21.0",
     "hydra-core==1.4.0.dev1",
+    "omegaconf==2.4.0.dev4",
     "word2number",
     "regex",
     "peft==0.11.1",
     "rlinf_latex2sympy2",
+    "mcp",
 ]
 agentic-vllm = [
     "vllm==0.8.5",
+    "numpy==2.2",
     "transformers==4.51.1",
     # uvloop is not compatible with simulators, put it here
     "uvloop==0.21.0",
     "hydra-core==1.4.0.dev1",
+    "omegaconf==2.4.0.dev4",
     "word2number",
     "regex",
     "peft==0.11.1",
     "rlinf_latex2sympy2",
+    "mcp",
 ]
 embodied = [
-    "transformers==4.40.1",
-    "draccus",
-    "rich",
-    "tensorflow",
-    "tensorflow_graphics",
-    "peft==0.11.1",
-    "timm==0.9.10",
-    "sapien==3.0.1;platform_system=='Linux'",
-    "tensordict",
+    "transformers<=4.57.6",
+    "peft",
+    "timm",
     "imageio[ffmpeg]",
-    "robosuite==1.4.1",
-    "bddl",
-    "easydict",
-    "cloudpickle",
-    "gym",
     "gymnasium",
-    "av",
-    "open3d",
-    "zarr",
-    "openai",
-    "jax",
+    "gym",
 ]
 franka = [
     "pyyaml",
@@ -142,6 +142,9 @@ override-dependencies = [
     "xgrammar==0.1.19",
     "uvloop==0.21.0",
     "tensorflow-addons==0.23.0",
+    "setuptools>=69.5.1,<75.9",
+    "datasets==3.6.0",
+    "peft==0.11.1",
 ]
 
 [tool.setuptools]

{rlinf-0.2.0.dev1 → rlinf-0.2.0.dev2}/rlinf/envs/__init__.py

@@ -45,9 +45,14 @@ def get_env_cls(env_type: str, env_cfg=None):
     env_type = SupportedEnvType(env_type)
 
     if env_type == SupportedEnvType.MANISKILL:
-        from rlinf.envs.maniskill.maniskill_env import ManiskillEnv
+        if env_cfg.get("enable_offload", False):
+            from rlinf.envs.maniskill.maniskill_offload_env import ManiskillOffloadEnv
 
-        return ManiskillEnv
+            return ManiskillOffloadEnv
+        else:
+            from rlinf.envs.maniskill.maniskill_env import ManiskillEnv
+
+            return ManiskillEnv
     elif env_type == SupportedEnvType.LIBERO:
         from rlinf.envs.libero.libero_env import LiberoEnv