Video not rendering, error code, gsplat: No CUDA toolkit found. gsplat will be disabled.

[gituser123456789000]

(sharp) C:\Users\USERNAME\ml-sharp>sharp predict -i C:/Users/USERNAME/ml-sharp/input -o C:/Users/USERNAME/ml-sharp/output --render
2025-12-29 04:48:00,544 | INFO | Processing 38 valid image files.
2025-12-29 04:48:00,556 | INFO | Using device cuda
2025-12-29 04:48:00,556 | INFO | No checkpoint provided. Downloading default model from https://ml-site.cdn-apple.com/models/sharp/sharp_2572gikvuh.pt
2025-12-29 04:48:01,648 | INFO | Using preset ViT dinov2l16_384.
2025-12-29 04:48:04,541 | INFO | Using preset ViT dinov2l16_384.
2025-12-29 04:48:08,690 | INFO | Processing C:\Users\USERNAME\ml-sharp\input\FILENAME.jpeg
2025-12-29 04:48:08,692 | WARNING | Did not find focallength in exif data of C:\Users\USERNAME\ml-sharp\input\FILENAME.jpeg - Setting to 30mm.
2025-12-29 04:48:08,812 | INFO | Running preprocessing.
2025-12-29 04:48:08,905 | INFO | Running inference.
2025-12-29 04:48:09,925 | INFO | Running postprocessing.
2025-12-29 04:48:12,658 | WARNING | Received 3 reflection matrices from SVD. Flipping them to rotations.
2025-12-29 04:48:13,213 | INFO | Saving 3DGS to C:\Users\USERNAME\ml-sharp\output
2025-12-29 04:48:15,588 | INFO | Rendering trajectory to C:\Users\USERNAME\ml-sharp\output\FILENAME.mp4
gsplat: No CUDA toolkit found. gsplat will be disabled.
Traceback (most recent call last):
File "", line 198, in _run_module_as_main
File "", line 88, in run_code
File "C:\Users\USERNAME\miniconda3\envs\sharp\Scripts\sharp.exe_main.py", line 6, in
sys.exit(main_cli())
~~~~~~~~^^
File "C:\Users\USERNAME\miniconda3\envs\sharp\Lib\site-packages\click\core.py", line 1462, in call
return self.main(*args, **kwargs)
~~~~~~~~~^^^^^^^^^^^^^^^^^
File "C:\Users\USERNAME\miniconda3\envs\sharp\Lib\site-packages\click\core.py", line 1383, in main
rv = self.invoke(ctx)
File "C:\Users\USERNAME\miniconda3\envs\sharp\Lib\site-packages\click\core.py", line 1850, in invoke
return _process_result(sub_ctx.command.invoke(sub_ctx))
~~~~~~~~~~~~~~~~~~~~~~^^^^^^^^^
File "C:\Users\USERNAME\miniconda3\envs\sharp\Lib\site-packages\click\core.py", line 1246, in invoke
return ctx.invoke(self.callback, **ctx.params)
~~~~~~~~~~^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "C:\Users\USERNAME\miniconda3\envs\sharp\Lib\site-packages\click\core.py", line 814, in invoke
return callback(*args, **kwargs)
File "C:\Users\USERNAME\ml-sharp\src\sharp\cli\predict.py", line 155, in predict_cli
render_gaussians(gaussians, metadata, output_video_path)
~~~~~~~~~~~~~~~~^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "C:\Users\USERNAME\ml-sharp\src\sharp\cli\render.py", line 110, in render_gaussians
rendering_output = renderer(
gaussians.to(device),
...<3 lines>...
image_height=camera_info.height,
)
File "C:\Users\USERNAME\miniconda3\envs\sharp\Lib\site-packages\torch\nn\modules\module.py", line 1773, in _wrapped_call_impl
return self._call_impl(*args, **kwargs)
~~~~~~~~~~~~~~~^^^^^^^^^^^^^^^^^
File "C:\Users\USERNAME\miniconda3\envs\sharp\Lib\site-packages\torch\nn\modules\module.py", line 1784, in _call_impl
return forward_call(*args, **kwargs)
File "C:\Users\USERNAME\ml-sharp\src\sharp\utils\gsplat.py", line 93, in forward
colors, alphas, meta = gsplat.rendering.rasterization(
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~^
means=gaussians.mean_vectors[ib],
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
...<12 lines>...
eps2d=self.low_pass_filter_eps,
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
)
^
File "C:\Users\USERNAME\miniconda3\envs\sharp\Lib\site-packages\gsplat\rendering.py", line 409, in rasterization
proj_results = fully_fused_projection(
means,
...<15 lines>...
opacities=opacities, # use opacities to compute a tigher bound for radii.
)
File "C:\Users\USERNAME\miniconda3\envs\sharp\Lib\site-packages\gsplat\cuda_wrapper.py", line 423, in fully_fused_projection
return _FullyFusedProjection.apply(
~~~~~~~~~~~~~~~~~~~~~~~~~~~^
means,
^^^^^^
...<13 lines>...
opacities,
^^^^^^^^^^
)
^
File "C:\Users\USERNAME\miniconda3\envs\sharp\Lib\site-packages\torch\autograd\function.py", line 576, in apply
return super().apply(*args, **kwargs) # type: ignore[misc]
~~~~~~~~~~~~~^^^^^^^^^^^^^^^^^
File "C:\Users\USERNAME\miniconda3\envs\sharp\Lib\site-packages\gsplat\cuda_wrapper.py", line 1056, in forward
camera_model_type = _make_lazy_cuda_obj(
f"CameraModelType.{camera_model.upper()}"
)
File "C:\Users\USERNAME\miniconda3\envs\sharp\Lib\site-packages\gsplat\cuda_wrapper.py", line 28, in _make_lazy_cuda_obj obj = getattr(_C, name_split)
AttributeError: 'NoneType' object has no attribute 'CameraModelType'

I have 4090, CUDA Toolkit 13.1 installed, Graphic Drivers version 591.xx (I've read 580.xx+ is needed for CUDA Toolkit 13.1, so that should be fine).
Any ideas on fixing this?

[gituser123456789000]

Restarting PC moved past that error..

tried running again.. there's a limit of 10 .PLY in the folder to render at a time (an error message states this).. I had like 20, so I cut down and tried again...
Then got a new error:

(sharp) C:\Users\USERNAME\ml-sharp>sharp render -i C:/Users/USERNAME/ml-sharp/output -o C:/Users/USERNAME/ml-sharp/output/renderings
2025-12-29 06:11:45,785 | INFO | Rendering C:\Users\USERNAME\ml-sharp\output\FILENAME.ply
W1229 06:11:46.998000 8512 site-packages\torch\utils\cpp_extension.py:466] Error checking compiler version for cl: [WinError 2] The system cannot find the file specified
W1229 06:11:46.999000 8512 site-packages\torch\utils\cpp_extension.py:2425] TORCH_CUDA_ARCH_LIST is not set, all archs for visible cards are included for compilation.
W1229 06:11:46.999000 8512 site-packages\torch\utils\cpp_extension.py:2425] If this is not desired, please set os.environ['TORCH_CUDA_ARCH_LIST'] to specific architectures.
INFO: Could not find files for the given pattern(s).
Traceback (most recent call last):
File "C:\Users\USERNAME\miniconda3\envs\sharp\Lib\site-packages\gsplat\cuda_backend.py", line 166, in
from gsplat import csrc as C
ImportError: cannot import name 'csrc' from 'gsplat' (C:\Users\USERNAME\miniconda3\envs\sharp\Lib\site-packages\gsplat_init.py)

During handling of the above exception, another exception occurred:

Traceback (most recent call last):
File "C:\Users\USERNAME\miniconda3\envs\sharp\Lib\site-packages\gsplat\cuda_backend.py", line 79, in load_extension
compiled = _jit_compile(
name,
...<10 lines>...
keep_intermediates=True,
)
TypeError: _jit_compile() missing 2 required positional arguments: 'verbose' and 'with_sycl'

During handling of the above exception, another exception occurred:

Traceback (most recent call last):
File "", line 198, in _run_module_as_main
File "", line 88, in run_code
File "C:\Users\USERNAME\miniconda3\envs\sharp\Scripts\sharp.exe_main.py", line 6, in
sys.exit(main_cli())
~~~~~~~~^^
File "C:\Users\USERNAME\miniconda3\envs\sharp\Lib\site-packages\click\core.py", line 1462, in call
return self.main(*args, **kwargs)
~~~~~~~~~^^^^^^^^^^^^^^^^^
File "C:\Users\USERNAME\miniconda3\envs\sharp\Lib\site-packages\click\core.py", line 1383, in main
rv = self.invoke(ctx)
File "C:\Users\USERNAME\miniconda3\envs\sharp\Lib\site-packages\click\core.py", line 1850, in invoke
return _process_result(sub_ctx.command.invoke(sub_ctx))
~~~~~~~~~~~~~~~~~~~~~~^^^^^^^^^
File "C:\Users\USERNAME\miniconda3\envs\sharp\Lib\site-packages\click\core.py", line 1246, in invoke
return ctx.invoke(self.callback, **ctx.params)
~~~~~~~~~~^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "C:\Users\USERNAME\miniconda3\envs\sharp\Lib\site-packages\click\core.py", line 814, in invoke
return callback(*args, **kwargs)
File "C:\Users\USERNAME\ml-sharp\src\sharp\cli\render.py", line 62, in render_cli
render_gaussians(
~~~~~~~~~~~~~~~~^
gaussians=gaussians,
^^^^^^^^^^^^^^^^^^^^
...<2 lines>...
output_path=(output_path / scene_path.stem).with_suffix(".mp4"),
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
)
^
File "C:\Users\USERNAME\ml-sharp\src\sharp\cli\render.py", line 110, in render_gaussians
rendering_output = renderer(
gaussians.to(device),
...<3 lines>...
image_height=camera_info.height,
)
File "C:\Users\USERNAME\miniconda3\envs\sharp\Lib\site-packages\torch\nn\modules\module.py", line 1773, in _wrapped_call_impl
return self._call_impl(*args, **kwargs)
~~~~~~~~~~~~~~~^^^^^^^^^^^^^^^^^
File "C:\Users\USERNAME\miniconda3\envs\sharp\Lib\site-packages\torch\nn\modules\module.py", line 1784, in _call_impl
return forward_call(*args, **kwargs)
File "C:\Users\USERNAME\ml-sharp\src\sharp\utils\gsplat.py", line 93, in forward
colors, alphas, meta = gsplat.rendering.rasterization(
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~^
means=gaussians.mean_vectors[ib],
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
...<12 lines>...
eps2d=self.low_pass_filter_eps,
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
)
^
File "C:\Users\USERNAME\miniconda3\envs\sharp\Lib\site-packages\gsplat\rendering.py", line 409, in rasterization
proj_results = fully_fused_projection(
means,
...<15 lines>...
opacities=opacities, # use opacities to compute a tigher bound for radii.
)
File "C:\Users\USERNAME\miniconda3\envs\sharp\Lib\site-packages\gsplat\cuda_wrapper.py", line 423, in fully_fused_projection
return _FullyFusedProjection.apply(
~~~~~~~~~~~~~~~~~~~~~~~~~~~^
means,
^^^^^^
...<13 lines>...
opacities,
^^^^^^^^^^
)
^
File "C:\Users\USERNAME\miniconda3\envs\sharp\Lib\site-packages\torch\autograd\function.py", line 576, in apply
return super().apply(*args, **kwargs) # type: ignore[misc]
~~~~~~~~~~~~~^^^^^^^^^^^^^^^^^
File "C:\Users\USERNAME\miniconda3\envs\sharp\Lib\site-packages\gsplat\cuda_wrapper.py", line 1056, in forward
camera_model_type = _make_lazy_cuda_obj(
f"CameraModelType.{camera_model.upper()}"
)
File "C:\Users\USERNAME\miniconda3\envs\sharp\Lib\site-packages\gsplat\cuda_wrapper.py", line 24, in _make_lazy_cuda_obj
from ._backend import _C
File "C:\Users\USERNAME\miniconda3\envs\sharp\Lib\site-packages\gsplat\cuda_backend.py", line 217, in
_C = load_extension(
name=name,
...<5 lines>...
verbose=VERBOSE,
)
File "C:\Users\USERNAME\miniconda3\envs\sharp\Lib\site-packages\gsplat\cuda_backend.py", line 98, in load_extension
compiled = _jit_compile(
name,
...<12 lines>...
keep_intermediates=True,
)
File "C:\Users\USERNAME\miniconda3\envs\sharp\Lib\site-packages\torch\utils\cpp_extension.py", line 2138, in _jit_compile
_write_ninja_file_and_build_library(
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~^
name=name,
^^^^^^^^^^
...<9 lines>...
with_sycl=with_sycl,
^^^^^^^^^^^^^^^^^^^^
is_standalone=is_standalone)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "C:\Users\USERNAME\miniconda3\envs\sharp\Lib\site-packages\torch\utils\cpp_extension.py", line 2275, in _write_ninja_file_and_build_library
_write_ninja_file_to_build_library(
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~^
path=build_file_path,
^^^^^^^^^^^^^^^^^^^^^
...<8 lines>...
with_sycl=with_sycl,
^^^^^^^^^^^^^^^^^^^^
is_standalone=is_standalone)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "C:\Users\USERNAME\miniconda3\envs\sharp\Lib\site-packages\torch\utils\cpp_extension.py", line 2762, in _write_ninja_file_to_build_library
_write_ninja_file(
~~~~~~~~~~~~~~~~~^
path=path,
^^^^^^^^^^
...<12 lines>...
with_cuda=with_cuda,
^^^^^^^^^^^^^^^^^^^^
with_sycl=with_sycl)
^^^^^^^^^^^^^^^^^^^^
File "C:\Users\USERNAME\miniconda3\envs\sharp\Lib\site-packages\torch\utils\cpp_extension.py", line 2946, in _write_ninja_file
cl_paths = subprocess.check_output(['where',
~~~~~~~~~~~~~~~~~~~~~~~^^^^^^^^^^
'cl']).decode(*SUBPROCESS_DECODE_ARGS).split('\r\n')
^^^^^^
File "C:\Users\USERNAME\miniconda3\envs\sharp\Lib\subprocess.py", line 472, in check_output
return run(*popenargs, stdout=PIPE, timeout=timeout, check=True,
~~~^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
**kwargs).stdout
^^^^^^^^^
File "C:\Users\USERNAME\miniconda3\envs\sharp\Lib\subprocess.py", line 577, in run
raise CalledProcessError(retcode, process.args,
output=stdout, stderr=stderr)
subprocess.CalledProcessError: Command '['where', 'cl']' returned non-zero exit status 1.