🚴‍♂️ Installation

Dependencies

Python versions:

Python 3.8, 3.10

Operating systems:

Linux: Ubuntu 18.04+, Centos 7+

Hardware:

Rendering: NVIDIA or AMD GPU
Ray tracing: NVIDIA RTX GPU or AMD equivalent
Ray-tracing Denoising: NVIDIA GPU
GPU Simulation: NVIDIA GPU

Software:

Ray tracing: NVIDIA Driver >= 470
Denoising (OIDN): NVIDIA Driver >= 520

0. Install Vulkan

sudo apt install libvulkan1 mesa-vulkan-drivers vulkan-tools

1. Basic env

First, prepare a conda environment.

conda create -n RoboTwin python=3.8

conda activate RoboTwin

pip install torch==2.4.1 torchvision sapien==3.0.0b1 scipy==1.10.1 mplib==0.1.1 gymnasium==0.29.1 trimesh==4.4.3 open3d==0.18.0 imageio==2.34.2 pydantic openai huggingface_hub zarr

Then, install pytorch3d:

cd third_party/pytorch3d_simplified && pip install -e . && cd ../..

2. Assert download

''' python ./script/download_asset.py unzip aloha_urdf.zip && unzip gpt_models.zip '''

3. REMOVE !!!!!!!!!

3.1 Remove `convex=True`

You can use pip show mplib to find where the mplib installed.

# mplib.planner (mplib/planner.py) line 71
# remove `convex=True`

self.robot = ArticulatedModel(
            urdf,
            srdf,
            [0, 0, -9.81],
            user_link_names,
            user_joint_names,
            convex=True,
            verbose=False,
        )
=> 
self.robot = ArticulatedModel(
            urdf,
            srdf,
            [0, 0, -9.81],
            user_link_names,
            user_joint_names,
            # convex=True,
            verbose=False,
        )

3.2 Remove `or collide`

# mplib.planner (mplib/planner.py) line 848
# remove `or collide`

if np.linalg.norm(delta_twist) < 1e-4 or collide or not within_joint_limit:
                return {"status": "screw plan failed"}
=>
if np.linalg.norm(delta_twist) < 1e-4 or not within_joint_limit:
                return {"status": "screw plan failed"}

2.0 KiB Raw Blame History