nvidia isaacsim

相关链接备份5080系列架构不满足要求curobo是一个英伟达的运动规划库可以实现机器人规划官方文档https://curobo.org/get_started/1_install_instructions.htmlisaac sim是英伟达的模拟器文档官方链接https://docs.isaacsim.omniverse.nvidia.com/4.5.0/installation/install_workstation.html安装TurboVNC远程桌面安装curobo已在两台NVIDIA RTX 5000 Ada 服务器上基于cuda11.8完成配置请遵循下方操作并尽可能补充。一、下载与准备基于/ssd0/xiaojing下准备代码以本人账号 xiaojing 为例一curobogitclone https://github.com/NVlabs/curobo.git这一步拉代码会很慢请耐心等待。二Isaac-sim4.5https://docs.isaacsim.omniverse.nvidia.com/4.5.0/installation/download.html#isaac-sim-latest-release 下载4.5 linux 版本将下好的isaac-sim-standalone4.5.0-rc.36release.19112.f59b3005.gl.linux-x86_64.release.zip解压至isaacsim下mkdir/ssd0/xiaojing/isaacsimunzipisaac-sim-standalone4.5.0-rc.36release.19112.f59b3005.gl.linux-x86_64.release.zip-d/ssd0/xiaojing/isaacsim解压完可以直接打开cd/ssd0/xiaojing/isaacsim ./post_install.sh ./isaac-sim.selector.sh有问题参见https://docs.isaacsim.omniverse.nvidia.com/4.5.0/installation/install_workstation.html第一次运行需要等一会三Isaac Sim Assetshttps://docs.isaacsim.omniverse.nvidia.com/4.5.0/installation/download.html#isaac-sim-latest-release将下好的三个zip解压至isaacsim_assets下mkdir/ssd0/xiaojing/isaacsim_assetsunzipisaac-sim-assets-14.5.0-rc.36release.19112.f59b3005.zip-d/ssd0/xiaojing/isaacsim_assetsunzipisaac-sim-assets-24.5.0-rc.36release.19112.f59b3005.zip-d/ssd0/xiaojing/isaacsim_assetsunzipisaac-sim-assets-34.5.0-rc.36release.19112.f59b3005.zip-d/ssd0/xiaojing/isaacsim_assets在第二步解压好的issacsim目录下修改文件**apps/isaacsim.exp.base.kit**注意同时修改下方的路径[settings]persistent.isaac.asset_root.default/home/username/isaacsim_assets/Assets/Isaac/4.5exts.isaacsim.asset.browser.folders[/home/username/isaacsim_assets/Assets/Isaac/4.5/Isaac/Robots,/home/username/isaacsim_assets/Assets/Isaac/4.5/Isaac/People,/home/username/isaacsim_assets/Assets/Isaac/4.5/Isaac/IsaacLab,/home/username/isaacsim_assets/Assets/Isaac/4.5/Isaac/Props,/home/username/isaacsim_assets/Assets/Isaac/4.5/Isaac/Environments,/home/username/isaacsim_assets/Assets/Isaac/4.5/Isaac/Materials,/home/username/isaacsim_assets/Assets/Isaac/4.5/Isaac/Samples,/home/username/isaacsim_assets/Assets/Isaac/4.5/Isaac/Sensors,]然后一定要再打开一次issacsim注意修改路径:cd/ssd0/xiaojing/issacsim ./isaac-sim.sh --/persistent/isaac/asset_root/default/home/username/isaacsim_assets/Assets/Isaac/4.5这一步比较重要如果不能正常配置后续curobo就不能正常运行isaac-sim的部分。如果没运行这一步后续运行curobo部分会报错有疑问见https://docs.isaacsim.omniverse.nvidia.com/4.5.0/installation/install_faq.html 的Setup Tips部分。四nvblox选装是否安装取决于是否使用深度相机以及nvblox第一次安装请跳过此步等curobo和isaac sim完成后再回来。直接从二、修改~/.bashrc配置部分开始NVIDIA推出的nvblox是一个专门针对GPU优化的SDF库为用户提供了一系列高效的算法。如果不安装此步后续无法运行curobo中nvblox相关的demo例如omni_python examples/isaac_sim/motion_gen_reacher_nvblox.py报错为ile/ssd0/xiaojing/curobo/examples/isaac_sim/motion_gen_reacher_nvblox.py, line148,inmain2025-07-17 06:39:03[15,098ms][Error][omni.kit.app._impl][py stderr]: motion_gen_configMotionGenConfig.load_from_robot_config(2025-07-17 06:39:03[15,098ms][Error][omni.kit.app._impl][py stderr]: File/ssd0/xiaojing/curobo/src/curobo/wrap/reacher/motion_gen.py, line655,inload_from_robot_config2025-07-17 06:39:03[15,098ms][Error][omni.kit.app._impl][py stderr]: world_coll_checkercreate_collision_checker(world_cfg)2025-07-17 06:39:03[15,098ms][Error][omni.kit.app._impl][py stderr]: File/ssd0/xiaojing/curobo/src/curobo/geom/sdf/utils.py, line38,increate_collision_checker2025-07-17 06:39:03[15,098ms][Error][omni.kit.app._impl][py stderr]:returnWorldBloxCollision(config)2025-07-17 06:39:03[15,098ms][Error][omni.kit.app._impl][py stderr]: File/ssd0/xiaojing/curobo/src/curobo/geom/sdf/world_blox.py, line54,in__init__2025-07-17 06:39:03[15,098ms][Error][omni.kit.app._impl][py stderr]: super().__init__(config)2025-07-17 06:39:03[15,098ms][Error][omni.kit.app._impl][py stderr]: File/ssd0/xiaojing/curobo/src/curobo/geom/sdf/world_voxel.py, line39,in__init__2025-07-17 06:39:03[15,098ms][Error][omni.kit.app._impl][py stderr]: super().__init__(config)2025-07-17 06:39:03[15,098ms][Error][omni.kit.app._impl][py stderr]: File/ssd0/xiaojing/curobo/src/curobo/geom/sdf/world_mesh.py, line70,in__init__2025-07-17 06:39:03[15,098ms][Error][omni.kit.app._impl][py stderr]: super().__init__(config)2025-07-17 06:39:03[15,098ms][Error][omni.kit.app._impl][py stderr]: File/ssd0/xiaojing/curobo/src/curobo/geom/sdf/world.py, line722,in__init__2025-07-17 06:39:03[15,099ms][Error][omni.kit.app._impl][py stderr]: self.load_collision_model(self.world_model)2025-07-17 06:39:03[15,099ms][Error][omni.kit.app._impl][py stderr]: File/ssd0/xiaojing/curobo/src/curobo/geom/sdf/world_blox.py, line83,inload_collision_model2025-07-17 06:39:03[15,099ms][Error][omni.kit.app._impl][py stderr]: self._blox_mapperMapper(2025-07-17 06:39:03[15,099ms][Error][omni.kit.app._impl][py stderr]: TypeError: ABC()takes no arguments安装细节步骤参见https://curobo.org/get_started/1_install_instructions.html#nvblox-isaac-sim-install报错一第7步Install nvblox的时候可能会遇到问题/usr/bin/ld: /home/songwoo/pkgs/glog/install//lib/libglog.a(utilities.cc.o): in function google::GetStackTrace(void**, int, int): utilities.cc:(.text0x9e5): undefined reference to _Ux86_64_getcontext /usr/bin/ld: utilities.cc:(.text0x9fe): undefined reference to _ULx86_64_init_local /usr/bin/ld: utilities.cc:(.text0xa62): undefined reference to _ULx86_64_get_reg /usr/bin/ld: utilities.cc:(.text0xabb): undefined reference to _ULx86_64_step collect2: error: ld returned 1 exit status make[2]: *** [CMakeFiles/sphere_benchmark.dir/build.make:105: sphere_benchmark] error 1 make[1]: *** [CMakeFiles/Makefile2:318: CMakeFiles/sphere_benchmark.dir/all] error 2 make[1]: *** waiting for the unfinished work....解决方法见https://github.com/NVlabs/curobo/issues/180sudoaptremove libunwind-devDelete glog build folder, recreate, cmake, make, make installDelete nvblox build folder, recreate, cmake, make, make installrobby8服务器已经适配确认可行报错二 第8步nvblox_torch 可能会报错– Checking for module ‘libglog’ – No package ‘libglog’ found做法注意修改路径exportPKG_CONFIG_PATH/ssd0/xiaojing/pkgs/glog/install/lib/pkgconfig:$PKG_CONFIG_PATH报错三第8步nvblox_torch 可能会报错/usr/bin/ld: cannot find -lglog: No such file or directory做法将/ssd0/xiaojing/pkgs/nvblox_torch/src/nvblox_torch/cpp/CMakeLists.txt 替换为如下内容注意set(GLOG_DIR “/ssd0/xiaojing/pkgs/glog/install”) 这里请替换为自己的glog路径#[[Copyright(c)2023NVIDIA CORPORATIONAFFILIATES. All rights reserved. NVIDIA CORPORATION, its affiliates and licensors retain all intellectual property and proprietary rightsinand to this material, related documentation and any modifications thereto. Any use, reproduction, disclosure or distribution of this material and related documentation without an express license agreement from NVIDIA CORPORATION or its affiliates is strictly prohibited.]]cmake_minimum_required(VERSION3.16FATAL_ERROR)project(nvblox_torch_cpp LANGUAGES CXX CUDA)set(CMAKE_BUILD_TYPE RelWithDebInfo)set(CMAKE_CXX_STANDARD17)# pytorch 2.0 requires 17 while pytorch 1.x works with 14.set(CMAKE_CXX_STANDARD_REQUIRED True)if(CMAKE_VERSION VERSION_GREATER_EQUAL3.18)set(CUDA_ARCHITECTURE_FLAGS70)else()set(CUDA_ARCHITECTURE_FLAGS-gencodearchcompute_70,codesm_70 )endif()option(BUILD_TESTINGBuild testsOFF)# This asks include files from nvblox to not provide implementations# that cause linking against ABI-incompatible stuffADD_DEFINITIONS(-DPRE_CXX11_ABI_LINKABLE)SET(nvblox_link_options-Wl,--disable-new-dtags)find_package(Torch REQUIRED)set(CMAKE_CXX_FLAGS${CMAKE_CXX_FLAGS}${TORCH_CXX_FLAGS})find_package(PkgConfig REQUIRED)#pkg_check_modules(glog REQUIRED libglog)# Download thirdparty deps#message(STATUS Downloading 3rdparty dependencies)#message(STATUS Downloading Eigen)include(thirdparty/eigen/eigen.cmake)#find_package(Glog REQUIRED)#find_package(Eigen3 REQUIRED) # we need to import eigen from nvblox?set(GLOG_DIR/ssd0/xiaojing/pkgs/glog/install)include_directories(${GLOG_DIR}/include)find_library(GLOG_LIB NAMES glog PATHS${GLOG_DIR}/lib NO_DEFAULT_PATH)find_package(CUDA REQUIRED)find_package(nvblox REQUIRED)# compile cuda library:# Pass options to NVCCset(CMAKE_CUDA_FLAGS${CMAKE_CUDA_FLAGS}--expt-relaxed-constexpr --disable-warnings --generate-line-info -lineinfo -Xcudafe --display_error_number)set(CUDA_NVCC_FLAGS${CUDA_NVCC_FLAGS}--compiler-options -fPIC -stdc14)############# INCLUDES ############## Internal includesinclude_directories(include)# External includesinclude_directories(${CMAKE_CUDA_TOOLKIT_INCLUDE_DIRECTORIES})#include_directories(${GLOG_INCLUDE_DIRS} ${gflags_INCLUDE_DIRS})include_directories(${gflags_INCLUDE_DIRS})############## LIBRARIES ################# Nvblox-torch python wrapperadd_library(py_nvblox SHARED src/py_nvblox.cu src/py_mapper.cu src/py_scene.cpp src/convert_tensors.cpp src/sdf_query.cu src/sdf_cost_query.cu)target_link_libraries(py_nvblox nvblox_lib nvblox_eigen#${glog_LIBRARIES}${GLOG_LIB}${TORCH_LIBRARIES}${CUDA_LIBRARIES}#${CUDA_nv}#${CUDA_nvToolsExt_LIBRARY})set_target_properties(py_nvblox PROPERTIES CUDA_SEPARABLE_COMPILATION ON)########## TESTS ##########include(CTest)if(BUILD_TESTING)find_package(GTest REQUIRED)enable_testing()add_subdirectory(tests)endif()然后再执行https://curobo.org/get_started/1_install_instructions.html#nvblox-isaac-sim-install 中的第8步继续往下shinstall.sh$(omni_python-cimport torch.utils; print(torch.utils.cmake_prefix_path))报错四第9步python -m pip install opencv-python pyrealsense2 transforms3d 可能是有问题的它会把numpy1.24.0删掉装上最新版本这会导致提示scipy 1.11.4 requires numpy1.28.0,1.21.6, but you have numpy 2.2.6 which is incompatible.numba 0.59.1 requires numpy1.27,1.22, but you have numpy 2.2.6 which is incompatible.所以这里需要改下装合适的opencv版本omni_python-mpipinstallopencv-python4.8.0.74报错五在一切安装完毕以后在实际运行nvblox相关的任务时例如omni_python examples/isaac_sim/motion_gen_reacher_nvblox.py可能会报错OSError: libnvblox_lib.so: cannot open shared object file: No such file or directory在当前终端需要输入或写入 gedit ~/.bashrc中exportLD_LIBRARY_PATH/usr/local/lib再次运行即可正常显示此外如果想运行这些样例omni_python realsense_reacher.py可能会报错2025-07-18 07:12:08 [12,094ms] [Error] [omni.kit.app._impl] [py stderr]: from omni.isaac.core.materials.deformable_material import DeformableMaterial, DeformableMaterialView2025-07-18 07:12:08 [12,094ms] [Error] [omni.kit.app._impl] [py stderr]: ImportError: cannot import name ‘DeformableMaterialView’ from ‘omni.isaac.core.materials.deformable_material’ (/ssd0/xiaojing/isaacsim/extsDeprecated/omni.isaac.core/omni/isaac/core/materials/deformable_material.py)请打开/ssd0/xiaojing/isaacsim/extsDeprecated/omni.isaac.core/omni/isaac/core/materials/init.py这个文件将内容修改为注意第9行和第10行# Copyright (c) 2021-2024, NVIDIA CORPORATION. All rights reserved.## NVIDIA CORPORATION and its licensors retain all intellectual property# and proprietary rights in and to this software, related documentation# and any modifications thereto. Any use, reproduction, disclosure or# distribution of this software and related documentation without an express# license agreement from NVIDIA CORPORATION is strictly prohibited.#from omni.isaac.core.materials.deformable_materialimportDeformableMaterial from omni.isaac.core.materials.deformable_material_viewimportDeformableMaterialView from omni.isaac.core.materials.omni_glassimportOmniGlass from omni.isaac.core.materials.omni_pbrimportOmniPBR from omni.isaac.core.materials.particle_materialimportParticleMaterial from omni.isaac.core.materials.particle_material_viewimportParticleMaterialView from omni.isaac.core.materials.physics_materialimportPhysicsMaterial from omni.isaac.core.materials.preview_surfaceimportPreviewSurface from omni.isaac.core.materials.visual_materialimportVisualMaterial这里看起来是英伟达自己的bugDeformableMaterialView 定义在deformable_material_view.py文件中而不是deformable_material.py文件中。二、修改~/.bashrc配置#使用11.8版本cudaexportCUDA_HOME/usr/local/cuda-11.8exportPATH/usr/local/cuda-11.8/bin:$PATHexportLD_LIBRARY_PATH/usr/local/cuda-11.8/lib64:$LD_LIBRARY_PATH#和远程桌面可视化有关exportMESA_GL_VERSION_OVERRIDE4.6export__GLX_VENDOR_LIBRARY_NAMEnvidia#参见https://developer.nvidia.com/cuda-gpus根据显卡型号选择编号exportTORCH_CUDA_ARCH_LIST8.9#使用isaacsim下的python注意路径配置这是一个满足要求的python3.10版本aliasomni_python/ssd0/xiaojing/isaacsim/python.sh如果没有中间两个GLX的export会报错应该仅在VNC虚拟桌面的情况下报错xiaojingrobby4-Precision-7960-Tower:/ssd0/xiaojing/isaacsim$ ./isaac-sim.selector.sh[Info] [carb] Logging to file: /home/xiaojing/.nvidia-omniverse/logs/Kit/Isaac-Sim_App_Selector/4.5/kit_20250714_134833.log[0.427s] [ext: omni.kit.async_engine-0.0.1] startup[0.473s] [ext: omni.client.lib-1.0.0] startup[0.505s] [ext: omni.stats-1.0.1] startup[0.507s] [ext: omni.client-1.2.2] startup[0.514s] [ext: omni.gpu_foundation.shadercache.vulkan-1.0.0] startup[0.525s] [ext: omni.assets.plugins-0.0.0] startup[0.526s] [ext: omni.gpu_foundation-0.0.0] startup[0.535s] [ext: carb.windowing.plugins-1.0.0] startup[0.545s] [ext: omni.kit.renderer.init-0.0.0] startupX Error of failed request: GLXBadFBConfigMajor opcode of failed request: 150 (GLX)Minor opcode of failed request: 0 ()Serial number of failed request: 224Current serial number in output stream: 224三、安装依赖omni_python-mpipinstalltomli wheel ninja omni_python-mpipinstall-force-reinstalltorch2.5.1torchvision0.20.1 --index-url https://download.pytorch.org/whl/cu118cd/ssd0/xiaojing/curobo omni_python-mpipinstall-e.[isaacsim]--no-build-isolation注一、如果omni_python -m pip install tomli wheel ninja 这一步没有完成可能会卡在omni_python -m pip install -e .[isaacsim] --no-build-isolation 这里。二、可能有这些报错ERROR: pip’s dependency resolver does not currently take into account all the packages that are installed. This behaviour is the source of the following dependency conflicts.nvidia-srl-usd 0.19.1 requires usd-core24.00,21.11, which is not installed.nvidia-srl-usd-to-urdf 0.7.2 requires usd-core24.00,21.11, which is not installed.nvidia-srl-usd-to-urdf 0.7.2 requires lxml5.0.0,4.9.2, but you have lxml 5.3.0 which is incompatible.osqp 0.6.5 requires scipy1.12.0,0.13.2, but you have scipy 1.15.3 which is incompatible.尝试安装依赖omni_python-mpipinstalllxml4.9.2 omni_python-mpipinstallusd-core23.2omni_python-mpipinstallnumpy1.24omni_python-mpipinstallscipy1.11.4注意这几个包不要手动去装这是通过isaacsim自己带的其版本omni_python -m pip list需要是nvidia_srl_base 0.10.0nvidia-srl-math 0.14.0nvidia-srl-usd 0.19.1nvidia-srl-usd-to-urdf 0.7.2如果手动装了版本会变成这样 nvidia-srl-base 1.5.0nvidia-srl-math 1.1.0nvidia-srl-usd 1.0.0nvidia-srl-usd-to-urdf 1.0.1版本不对会报错再次在curobo下打开一个样例程序omni_python examples/isaac_sim/motion_gen_reacher.py成功启动常见报错一class ‘RuntimeError’ Used null prim2025-07-15 03:01:15 [12,627ms] [Error] [omni.kit.app._impl] [py stderr]: Traceback (most recent call last):2025-07-15 03:01:15 [12,627ms] [Error] [omni.kit.app._impl] [py stderr]: File “/ssd0/xiaojing/curobo/examples/isaac_sim/motion_gen_reacher.py”, line 446, in2025-07-15 03:01:15 [12,627ms] [Error] [omni.kit.app._impl] [py stderr]: main()2025-07-15 03:01:15 [12,627ms] [Error] [omni.kit.app._impl] [py stderr]: File “/ssd0/xiaojing/curobo/examples/isaac_sim/motion_gen_reacher.py”, line 190, in main2025-07-15 03:01:15 [12,628ms] [Error] [omni.kit.app._impl] [py stderr]: robot, robot_prim_path add_robot_to_scene(robot_cfg, my_world)2025-07-15 03:01:15 [12,628ms] [Error] [omni.kit.app._impl] [py stderr]: File “/ssd0/xiaojing/curobo/examples/isaac_sim/helper.py”, line 141, in add_robot_to_scene2025-07-15 03:01:15 [12,628ms] [Error] [omni.kit.app._impl] [py stderr]: str(my_world.scene.stage.GetDefaultPrim().GetPath()) robot_path,2025-07-15 03:01:15 [12,628ms] [Error] [omni.kit.app._impl] [py stderr]: TypeError: can only concatenate str (not “NoneType”) to str两种原因第一种pip源版本不对这是我的可以跑的版本供参考aiobotocore2.12.1 aiodns3.1.1 aiofiles23.2.1 aiohttp3.9.4 aioitertools0.11.0 aiosignal1.3.1 annotated-types0.7.0 anyio4.6.2.post1 asteval0.9.32 async-timeout4.0.3 attrs23.2.0 azure-core1.28.0 azure-identity1.13.0 azure-storage-blob12.17.0 boto31.26.63 botocore1.34.68 cchardet2.1.7 certifi2024.8.30 cffi1.17.1 charset-normalizer3.3.2 click8.1.7 colorlog6.9.0 construct2.10.68 coverage7.4.4 cryptography44.0.0 cycler0.11.0 embreex2.17.7.post6 exceptiongroup1.2.2 fastapi0.110.0 filelock3.18.0 frozenlist1.4.1 fsspec2025.5.1 gunicorn22.0.0 gymnasium0.29.1 h110.16.0 httpcore1.0.9 httptools0.6.1 httpx0.28.1 idna3.6idna-ssl1.1.0 imageio2.37.0 importlib_resources6.5.2 isodate0.6.1 Jinja23.1.4 jmespath1.0.1 jsonschema4.24.0 jsonschema-specifications2025.4.1 kiwisolver1.4.4 lazy_loader0.4llvmlite0.42.0 lxml4.9.2 manifold3d3.1.1 mapbox_earcut1.0.3 MarkupSafe2.1.3 matplotlib3.8.4 mpmath1.3.0 msal1.23.0 msal-extensions1.0.0 multidict6.0.5 nest-asyncio1.5.6 networkx3.4.2 ninja1.11.1.4 numba0.59.1 numpy1.24.0 numpy-quaternion2023.0.3 nvidia-cublas-cu1111.11.3.6 nvidia-cuda-cupti-cu1111.8.87 nvidia-cuda-nvrtc-cu1111.8.89 nvidia-cuda-runtime-cu1111.8.89 nvidia-cudnn-cu119.1.0.70 nvidia-cufft-cu1110.9.0.58 nvidia-curand-cu1110.3.0.86 nvidia_curobo0.7.7.post1.dev2dirty /home/xiaojing/curobo nvidia-cusolver-cu1111.4.1.48 nvidia-cusparse-cu1111.7.5.86 nvidia-lula-no-cuda0.10.1 nvidia-nccl-cu112.21.5 nvidia-nvtx-cu1111.8.86 nvidia_srl_base0.10.0 nvidia-srl-math0.14.0 nvidia-srl-usd0.19.1 nvidia-srl-usd-to-urdf0.7.2 nvsmi0.4.2 oauthlib3.2.2 opencv-python-headless4.9.0.80 osqp0.6.5 packaging23.0pillow11.3.0 Pint0.20.1 pip24.0nv1 plotly5.3.1 portalocker2.7.0 psutil5.9.8 pybind113.0.0 pycares4.3.0 pycollada0.9.2 pycparser2.22pydantic2.9.2 pydantic_core2.23.4 pyparsing3.0.9 pyperclip1.8.0 pypng0.20220715.0 python-dateutil2.9.0.post0 python-multipart0.0.9 pytz2024.1PyYAML6.0.1 qdldl0.1.7.post1 qrcode7.4.2 referencing0.36.2 requests2.32.3 requests-oauthlib1.3.1 rpds-py0.26.0 rtree1.4.0 s3transfer0.6.1 scikit-image0.25.2 scipy1.11.4 selenium4.14.0 sentry-sdk1.43.0 setuptools70.3.0 setuptools-scm8.0.4 shapely2.1.1 six1.16.0 sniffio1.3.1 starlette0.36.3 svg.path7.0sympy1.13.1 tifffile2025.5.10 toml0.10.2 tomli2.0.1 torch2.5.1cu118 torchaudio2.5.0cu118 torchvision0.20.1cu118 tornado6.4.2 tqdm4.67.1 trimesh4.7.0 triton3.1.0 typing_extensions4.10.0 urllib32.2.3 usd-core23.2uvicorn0.29.0 vhacdx0.0.8.post2 warp-lang1.8.0 watchdog4.0.0 webbot0.34websockets12.0wheel0.45.1 wrapt1.16.0 xxhash3.5.0 yarl1.9.4 yourdfpy0.0.58第二种curobo下文件缺失尚不清楚是哪一步导致这些文件被删这些内容在运行时需要有。请通过git restore . 指令恢复原状常见报错二‘NoneType’ object has no attribute ‘GetPath’这个请看上方三部分Isaac Sim Assets部分的配置。