ImportError: /root/.cache/torch_extensions/py310_cu124/nvdiffrast_plugin/nvdiffrast_plugin.so: cannot open shared object file: No such file or directory

[FurkanGozukara]

Main issue : Tencent/Hunyuan3D-1#5

How I installed with no errors on Cuda 12.4, Python 3.10.11 : pip install git+https://github.com/NVlabs/nvdiffrast

RTX A6000 , ubuntu 22

100%|█████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████| 25/25 [00:15<00:00,  1.58it/s]
[HunYuan3D]-[text to image], cost time: 16.2816s
[HunYuan3D]-[remove background], cost time: 3.9235s
100%|█████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████| 50/50 [00:12<00:00,  4.09it/s]
[HunYuan3D]-[image to views], cost time: 13.0661s
./outputs/app_output/2
=====> LRM forward time: 2.2705509662628174
=====> generate mesh with vertex shading time: 2.4042084217071533
Using xatlas to perform UV unwrapping, may take a while ...
Traceback (most recent call last):
  File "/workspace/Hunyuan3D-1/venv/lib/python3.10/site-packages/gradio/queueing.py", line 624, in process_events
    response = await route_utils.call_process_api(
  File "/workspace/Hunyuan3D-1/venv/lib/python3.10/site-packages/gradio/route_utils.py", line 323, in call_process_api
    output = await app.get_blocks().process_api(
  File "/workspace/Hunyuan3D-1/venv/lib/python3.10/site-packages/gradio/blocks.py", line 2015, in process_api
    result = await self.call_function(
  File "/workspace/Hunyuan3D-1/venv/lib/python3.10/site-packages/gradio/blocks.py", line 1562, in call_function
    prediction = await anyio.to_thread.run_sync(  # type: ignore
  File "/workspace/Hunyuan3D-1/venv/lib/python3.10/site-packages/anyio/to_thread.py", line 56, in run_sync
    return await get_async_backend().run_sync_in_worker_thread(
  File "/workspace/Hunyuan3D-1/venv/lib/python3.10/site-packages/anyio/_backends/_asyncio.py", line 2441, in run_sync_in_worker_thread
    return await future
  File "/workspace/Hunyuan3D-1/venv/lib/python3.10/site-packages/anyio/_backends/_asyncio.py", line 943, in run
    result = context.run(func, *args)
  File "/workspace/Hunyuan3D-1/venv/lib/python3.10/site-packages/gradio/utils.py", line 865, in wrapper
    response = f(*args, **kwargs)
  File "/workspace/Hunyuan3D-1/app.py", line 182, in stage_3_v23
    worker_v23(
  File "/workspace/Hunyuan3D-1/venv/lib/python3.10/site-packages/torch/utils/_contextlib.py", line 116, in decorate_context
    return func(*args, **kwargs)
  File "/workspace/Hunyuan3D-1/infer/utils.py", line 49, in wrapper
    result = func(*args, **kwargs)
  File "/workspace/Hunyuan3D-1/infer/utils.py", line 66, in wrapper
    output = func(*args, **kwargs)
  File "/workspace/Hunyuan3D-1/infer/views_to_mesh.py", line 85, in __call__
    self.mv23d_predictor.predict(
  File "/workspace/Hunyuan3D-1/venv/lib/python3.10/site-packages/torch/utils/_contextlib.py", line 116, in decorate_context
    return func(*args, **kwargs)
  File "/workspace/Hunyuan3D-1/svrm/predictor.py", line 145, in predict
    self.model.export_mesh_with_uv(
  File "/workspace/Hunyuan3D-1/venv/lib/python3.10/site-packages/torch/utils/_contextlib.py", line 116, in decorate_context
    return func(*args, **kwargs)
  File "/workspace/Hunyuan3D-1/svrm/ldm/models/svrm.py", line 213, in export_mesh_with_uv
    ctx = NVDiffRasterizerContext(context_type, cur_triplane.device) if ctx is None else ctx
  File "/workspace/Hunyuan3D-1/svrm/ldm/modules/rendering_neus/rasterize.py", line 10, in __init__
    self.ctx = self.initialize_context(context_type, device)
  File "/workspace/Hunyuan3D-1/svrm/ldm/modules/rendering_neus/rasterize.py", line 18, in initialize_context
    return dr.RasterizeCudaContext(device=device)
  File "/workspace/Hunyuan3D-1/venv/lib/python3.10/site-packages/nvdiffrast/torch/ops.py", line 184, in __init__
    self.cpp_wrapper = _get_plugin().RasterizeCRStateWrapper(cuda_device_idx)
  File "/workspace/Hunyuan3D-1/venv/lib/python3.10/site-packages/nvdiffrast/torch/ops.py", line 125, in _get_plugin
    torch.utils.cpp_extension.load(name=plugin_name, sources=source_paths, extra_cflags=common_opts+cc_opts, extra_cuda_cflags=common_opts+['-lineinfo'], extra_ldflags=ldflags, with_cuda=True, verbose=False)
  File "/workspace/Hunyuan3D-1/venv/lib/python3.10/site-packages/torch/utils/cpp_extension.py", line 1314, in load
    return _jit_compile(
  File "/workspace/Hunyuan3D-1/venv/lib/python3.10/site-packages/torch/utils/cpp_extension.py", line 1746, in _jit_compile
    return _import_module_from_library(name, build_directory, is_python_module)
  File "/workspace/Hunyuan3D-1/venv/lib/python3.10/site-packages/torch/utils/cpp_extension.py", line 2140, in _import_module_from_library
    module = importlib.util.module_from_spec(spec)
  File "<frozen importlib._bootstrap>", line 571, in module_from_spec
  File "<frozen importlib._bootstrap_external>", line 1176, in create_module
  File "<frozen importlib._bootstrap>", line 241, in _call_with_frames_removed
ImportError: /root/.cache/torch_extensions/py310_cu124/nvdiffrast_plugin/nvdiffrast_plugin.so: cannot open shared object file: No such file or directory

[zbclovehj]

You can find the nvdiffrast_plugin, and to use ninja compile file.

[zbclovehj]

cd C:\Users*\AppData\Local\torch_extensions\torch_extensions\Cache\py311_cu124\nvdiffrast_plugin
ninja

[FurkanGozukara]

cd C:\Users*\AppData\Local\torch_extensions\torch_extensions\Cache\py311_cu124\nvdiffrast_plugin ninja

thanks a lot for answer will try

[jdc1202]

I have the same question. Are you fix the bug?

[devrosx]

hello i have same problem...
and ninja failed to compile....
`(env) pc@pc-desktop:~/pinokio/cache/XDG_CACHE_HOME/torch_extensions/py310_cu121/nvdiffrast_plugin$ ninja
[1/6] /home/pc/pinokio/bin/miniconda/bin/nvcc --generate-dependencies-with-compile --dependency-o...ib/python3.10/site-packages/nvdiffrast/common/cudaraster/impl/RasterImpl.cu -o RasterImpl.cuda.o
FAILED: RasterImpl.cuda.o
/home/pc/pinokio/bin/miniconda/bin/nvcc --generate-dependencies-with-compile --dependency-output RasterImpl.cuda.o.d -DTORCH_EXTENSION_NAME=nvdiffrast_plugin -DTORCH_API_INCLUDE_EXTENSION_H -DPYBIND11_COMPILER_TYPE="gcc" -DPYBIND11_STDLIB="libstdcpp" -DPYBIND11_BUILD_ABI="cxxabi1011" -isystem /home/pc/pinokio/api/comfy.git/app/env/lib/python3.10/site-packages/torch/include -isystem /home/pc/pinokio/api/comfy.git/app/env/lib/python3.10/site-packages/torch/include/torch/csrc/api/include -isystem /home/pc/pinokio/api/comfy.git/app/env/lib/python3.10/site-packages/torch/include/TH -isystem /home/pc/pinokio/api/comfy.git/app/env/lib/python3.10/site-packages/torch/include/THC -isystem /home/pc/pinokio/bin/miniconda/include -isystem /home/pc/pinokio/bin/miniconda/include/python3.10 -D_GLIBCXX_USE_CXX11_ABI=0 -D__CUDA_NO_HALF_OPERATORS -D__CUDA_NO_HALF_CONVERSIONS_ -D__CUDA_NO_BFLOAT16_CONVERSIONS__ -D__CUDA_NO_HALF2_OPERATORS__ --expt-relaxed-constexpr -gencode=arch=compute_86,code=compute_86 -gencode=arch=compute_86,code=sm_86 --compiler-options '-fPIC' -DNVDR_TORCH -lineinfo -std=c++17 -c /home/pc/pinokio/api/comfy.git/app/env/lib/python3.10/site-packages/nvdiffrast/common/cudaraster/impl/RasterImpl.cu -o RasterImpl.cuda.o
In file included from /home/pc/pinokio/bin/miniconda/include/cuda_runtime.h:83,
from :
/home/pc/pinokio/bin/miniconda/include/crt/host_config.h:132:2: error: #error -- unsupported GNU version! gcc versions later than 12 are not supported! The nvcc flag '-allow-unsupported-compiler' can be used to override this version check; however, using an unsupported host compiler may cause compilation failure or incorrect run time execution. Use at your own risk.
132 | #error -- unsupported GNU version! gcc versions later than 12 are not supported! The nvcc flag '-allow-unsupported-compiler' can be used to override this version check; however, using an unsupported host compiler may cause compilation failure or incorrect run time execution. Use at your own risk.
``

[zbclovehj]

Please translate this file into English. I hope the file can help you!

On Fri, Dec 27, 2024 at 3:53 AM devrosx ***@***.***> wrote: hello i have same problem... and ninja failed to compile.... `(env) ***@***.***:~/pinokio/cache/XDG_CACHE_HOME/torch_extensions/py310_cu121/nvdiffrast_plugin$ ninja [1/6] /home/pc/pinokio/bin/miniconda/bin/nvcc --generate-dependencies-with-compile --dependency-o...ib/python3.10/site-packages/nvdiffrast/common/cudaraster/impl/RasterImpl.cu -o RasterImpl.cuda.o FAILED: RasterImpl.cuda.o /home/pc/pinokio/bin/miniconda/bin/nvcc --generate-dependencies-with-compile --dependency-output RasterImpl.cuda.o.d -DTORCH_EXTENSION_NAME=nvdiffrast_plugin -DTORCH_API_INCLUDE_EXTENSION_H -DPYBIND11_COMPILER_TYPE="gcc" -DPYBIND11_STDLIB="libstdcpp" -DPYBIND11_BUILD_ABI="cxxabi1011" -isystem /home/pc/pinokio/api/comfy.git/app/env/lib/python3.10/site-packages/torch/include -isystem /home/pc/pinokio/api/comfy.git/app/env/lib/python3.10/site-packages/torch/include/torch/csrc/api/include -isystem /home/pc/pinokio/api/comfy.git/app/env/lib/python3.10/site-packages/torch/include/TH -isystem /home/pc/pinokio/api/comfy.git/app/env/lib/python3.10/site-packages/torch/include/THC -isystem /home/pc/pinokio/bin/miniconda/include -isystem /home/pc/pinokio/bin/miniconda/include/python3.10 -D_GLIBCXX_USE_CXX11_ABI=0 -D__CUDA_NO_HALF_OPERATORS -D__CUDA_NO_HALF_CONVERSIONS_ -D__CUDA_NO_BFLOAT16_CONVERSIONS__ -D__CUDA_NO_HALF2_OPERATORS__ --expt-relaxed-constexpr -gencode=arch=compute_86,code=compute_86 -gencode=arch=compute_86,code=sm_86 --compiler-options '-fPIC' -DNVDR_TORCH -lineinfo -std=c++17 -c /home/pc/pinokio/api/comfy.git/app/env/lib/python3.10/site-packages/nvdiffrast/common/cudaraster/impl/RasterImpl.cu -o RasterImpl.cuda.o In file included from /home/pc/pinokio/bin/miniconda/include/cuda_runtime.h:83, from : /home/pc/pinokio/bin/miniconda/include/crt/host_config.h:132:2: error: #error -- unsupported GNU version! gcc versions later than 12 are not supported! The nvcc flag '-allow-unsupported-compiler' can be used to override this version check; however, using an unsupported host compiler may cause compilation failure or incorrect run time execution. Use at your own risk. 132 | #error -- unsupported GNU version! gcc versions later than 12 are not supported! The nvcc flag '-allow-unsupported-compiler' can be used to override this version check; however, using an unsupported host compiler may cause compilation failure or incorrect run time execution. Use at your own risk. `` — Reply to this email directly, view it on GitHub <#207 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/BJN2XRD6KMUOY4OX5E3I3I32HRNEPAVCNFSM6AAAAABRGHH6ZCVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDKNRTGA3DAOJUGA> . You are receiving this because you commented.Message ID: ***@***.***>

ImportError: /home/data2/chuan/.conda/envs/DiffTex_New/lib/python3.10/site-packages/torch/lib/../../../.././ libstdc++.so.6: version `GLIBCXX_3.4.29' not found (required by /home/data2/chuan/.cache/torch_extensions/py310_cu118/nvdiffrast_plugin/nvdiffrast_plugin.so) 这个错误使用以下命令 rm /home/data2/chuan/.conda/envs/DiffTex_New/lib/libstdc++.so.6 cp /usr/lib/x86_64-linux-gnu/libstdc++.so.6.0.29 /home/data2/chuan/.conda/envs/DiffTex_New/lib ln -s /home/data2/chuan/.conda/envs/DiffTex_New/lib/libstdc++.so.6.0.29 /home/data2/chuan/.conda/envs/DiffTex_New/lib/libstdc++.so.6 nvcc fatal : Value ‘c++17’ is not defined for option ‘std’ 因为是第一次用这个机器，用户的环境变量里没有加新版本的cuda，于是在~/.bashrc里加入下面两行： export PATH="/usr/local/cuda-11.3/bin:$PATH" export LD_LIBRARY_PATH="/usr/local/cuda-11.3/lib64:$LD_LIBRARY_PATH" 我解决了这个问题： 1、首先，cl.exe ninja 的路径要对，要能找的到 2、你的 anaconda3\envs\instantmesh\Lib\site-packages\torch\utils\cpp_extension.py 中，将2065行改成 command = ['ninja', '--version']， 以解决ninja -v的错误 3、你的 anaconda3\envs\instantmesh\Lib\site-packages\nvdiffrast\torch\ops.py中，将118行改成 verbose=True 4、运行python app.py，大概聊仍然会报找不到nvdiffrast_plugin，这时候到 C:\Users****\AppData\Local\torch_extensions\Cache\py310_cu121\nvdiffrast_plugin 或者/home/data2/chuan/.cache/torch_extensions/py310_cu121/nvdiffrast_plugin 目录下，手动运行ninja，以生成nvdiffrast_plugin 注意，我这里还有一个问题，就是安装cuda12.1.0后，在系统环境变量中，没有自动生成CUDA_PATH环境变量，需要手动指向cuda12.1.0的目录。这也是为什么在ninja时，nvcc使用的是conda的nvcc，导致ninja编译失败！ rmdir /S %userprofile%\AppData\Local\torch_extensions\torch_extensions\Cache\py310_cu121\nvdiffrast_plugin set CUDA_PATH=C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v12.1\ computer89错误本质上是因为CUDA版本低所以不支持89的计算 把CUDA版本的环境变量换成高版本的CUDA路径即可 原因是conda环境下cuda编译器nvcc不支持compute_89导致安装失败， 运行项目的requirement.txt安装环境往往会默认安装低版本的cuda环境。 通过 nvcc --list-gpu-arch 来查看支持的编译器 如果没有compute_89（nvcc --version版本是11.5或者以下）， 可以修改.bashrc export PATH=/usr/local/cuda-12.1/bin:$PATH export LD_LIBRARY_PATH=/usr/local/cuda-12.1/lib64:$LD_LIBRARY_PATH 更新环境变量 source ~/.bashrc 再次查看nvcc --list-gpu-arch是否有compute_89 如果没有的话，可以用which nvcc查看nvcc目录 which nvcc 如果nvcc目录不是/usr/local/cuda-12.1 ，需要重新配置，先备份下原环境，然后重新配置 sudo mv /usr/bin/nvcc /usr/bin/nvcc.bak sudo ln -s /usr/local/cuda-12.1/bin/nvcc /usr/bin/nvcc 这样就可以了。 如果没有更高版本的cuda和nvcc就需要重新安装cuda和nvcc

[devrosx]

Thanks found way to fix it allready

MrForExample/ComfyUI-3D-Pack#411 (comment)