RoboTwin 2.0 Runtime for RLinf / RPent
RoboTwin 2.0 is a robot-manipulation simulation benchmark covering a wide range of single-arm and dual-arm tasks.
rlinf-robotwin-runtime is the RLinf-maintained RoboTwin environment
distribution used by RPent.
This package focuses on the simulation/runtime environment required by RPent. It is not intended to replace RoboTwin's full upstream development or training setup.
Installation
Standalone runtime
pip install rlinf-robotwin-runtime==0.1.2
The PyPI wheel contains the RoboTwin runtime code and the small task/config files required at runtime, but does not bundle the large simulation assets.
Download the required assets after installation:
robotwin-download-assets
RPent
For RPent users, install RoboTwin through RPent's robotwin extra rather than
assembling the environment and VLA dependencies manually:
uv pip install -e ".[robotwin]"
The RPent RoboTwin stack consists of:
RPent
├── rlinf-robotwin-runtime
├── rlinf-lingbotvla
└── NVIDIA cuRobo
Package Contents
The wheel contains the runtime-facing RoboTwin Python package, including:
- RoboTwin environment and robot runtime code
- task configuration YAML files
- task instruction / object description metadata
- the
robotwin-download-assetsCLI
Repository-level development utilities such as tests and helper scripts are not part of the runtime wheel.
Large simulator assets are also distributed separately.
Runtime Contract
The 0.1.2 runtime requires:
| Component | Version |
|---|---|
| Python | >=3.10,<3.12 |
| PyTorch | 2.8.0 |
| SAPIEN | 3.0.0b1 |
| mplib | 0.2.1 |
| Gymnasium | 0.29.1 |
| NumPy | 1.26.4 |
The dependency pins keep the simulator and VLA on the same Torch version when installed together through RPent.
cuRobo
cuRobo is intentionally not declared as a dependency of this PyPI package.
RoboTwin treats cuRobo as an optional planner dependency. RPent owns the source dependency for the complete integration and currently pins the official NVIDIA release:
NVlabs/curobo@v0.7.8
This keeps the public RoboTwin wheel free of direct Git dependencies while keeping the full RPent installation reproducible.
RLinf Compatibility
The RLinf-maintained source branch contains the environment/runtime integration used by RPent:
- Source: https://github.com/RLinf/RoboTwin
- Upstream RoboTwin: https://github.com/RoboTwin-Platform/RoboTwin
- RPent: https://github.com/RLinf/RPent
For the current list of tasks with RLinf-compatible reward support, see the README in the RLinf RoboTwin repository.
License
RoboTwin runtime code is distributed under the MIT License.
Please refer to the upstream RoboTwin project for additional dataset, asset, and third-party licensing information.
Metadata
Release files for rlinf-robotwin-runtime 0.1.2
For a detailed explanation of source distributions (sdists) and built distributions (wheels), please see the package formats documentation.
Source distribution (sdist)
| File | Size | Uploaded | |
|---|---|---|---|
| rlinf_robotwin_runtime-0.1.2.tar.gz | 224.9 kB | Details |
Built distribution (wheel)
| File | Interpreter | ABI | Platform | Reset |
|---|---|---|---|---|
| rlinf_robotwin_runtime-0.1.2-py3-none-any.whl | Python 3 | none | any | Details |
Total release size: 696.3 kB
Release files / rlinf_robotwin_runtime-0.1.2.tar.gz
| Download URL | rlinf_robotwin_runtime-0.1.2.tar.gz |
|---|---|
| Size | 224.9 kB |
| Tags | Source |
|
SHA-256 checksum How to use checksums |
c2d003e6ee532281c4c4d2f4560c7cd83afbd86d20bd65ded225092869fd71cb
|
|
BLAKE2b-256 checksum How to use checksums |
7079f6b3d25bcac36af00aab00a3889a0a94ca24ef069280b7c6c4a916d7a47f
|
| Upload date | |
|
Uploaded using Trusted Publishing? What is trusted publishing? |
No |
| Uploaded via |
twine/7.0.0 CPython/3.11.15
|
Release files / rlinf_robotwin_runtime-0.1.2-py3-none-any.whl
| Download URL | rlinf_robotwin_runtime-0.1.2-py3-none-any.whl |
|---|---|
| Size | 471.3 kB |
| Tags | Python 3 |
|
SHA-256 checksum How to use checksums |
2b1065befe7bb42172434f5c3c06f88f7ef9cbc6942f8ffa9115532ec95b008e
|
|
BLAKE2b-256 checksum How to use checksums |
78abae63f7faebe297654fd580d74dd1de5d8dc2185c097449d561fa73bb52db
|
| Upload date | |
|
Uploaded using Trusted Publishing? What is trusted publishing? |
No |
| Uploaded via |
twine/7.0.0 CPython/3.11.15
|