need help, compile successfully but cannot run the code

[Holmes-Alan]

I can successfully compile the codes but when I tried to train the model using

python main_nerf.py data/fox --workspace trial_nerf --fp16 --ff --cuda_ray --gui

I got the errors as,

Traceback (most recent call last):
File "main_nerf.py", line 47, in
from nerf.network_ff import NeRFNetwork
File "D:\torch-ngp\nerf\network_ff.py", line 5, in
from encoding import get_encoder
File "D:\torch-ngp\encoding.py", line 5, in
from hashencoder import HashEncoder
File "D:\torch-ngp\hashencoder_init_.py", line 1, in
from .hashgrid import HashEncoder
File "D:\torch-ngp\hashencoder\hashgrid.py", line 9, in
from .backend import _backend
File "D:\torch-ngp\hashencoder\backend.py", line 6, in
_backend = load(name='_hash_encoder',
File "C:\Users\Miniconda3\envs\ngp\lib\site-packages\torch\utils\cpp_extension.py", line 986, in load
return _jit_compile(
File "C:\Users\Miniconda3\envs\ngp\lib\site-packages\torch\utils\cpp_extension.py", line 1193, in _jit_compile
_write_ninja_file_and_build_library(
File "C:\Users\Miniconda3\envs\ngp\lib\site-packages\torch\utils\cpp_extension.py", line 1285, in _write_ninja_file_and_build_library
_write_ninja_file_to_build_library(
File "C:\Users\Miniconda3\envs\ngp\lib\site-packages\torch\utils\cpp_extension.py", line 1679, in _write_ninja_file_to_build_library
_write_ninja_file(
File "C:\Users\Miniconda3\envs\ngp\lib\site-packages\torch\utils\cpp_extension.py", line 1790, in _write_ninja_file
cl_paths = subprocess.check_output(['where',
File "C:\Users\Miniconda3\envs\ngp\lib\subprocess.py", line 415, in check_output
return run(*popenargs, stdout=PIPE, timeout=timeout, check=True,
File "C:\Users\Miniconda3\envs\ngp\lib\subprocess.py", line 516, in run
raise CalledProcessError(retcode, process.args,
subprocess.CalledProcessError: Command '['where', 'cl']' returned non-zero exit status 1.

I am using a computer with RTX3090, I tried using an environment with CUDA 11.0 and torch 1.7.2, I also tested on CUDA 11.3 and torch 1.9. They all failed.

Can you please tell me what is problem? Thank you!

[ashawkey]

Hi,
Are you using Windows or Linux? If it is on Windows, I found some related issues here and here, which seems to be a configuration problem of visual studio.

Besides, it's better to use torch 1.10, or you'll have to manually modify some lines (mostly removing indexing=... from torch.meshgrid() which does not exist in torch < 1.10)

[Holmes-Alan]

Thanks very much for the quick reply! I think I have solved the problem. Cheers!

[ashawkey]

Glad to hear that, closed for now.

[Holmes-Alan]

When I run

python main_nerf.py data/fox --workspace trial_nerf --fp16 --ff --cuda_ray --gui

I got the following error,

Traceback (most recent call last):
File "C:\Users\zhisong_liu\Miniconda3\envs\ngp\lib\site-packages\torch\utils\cpp_extension.py", line 1740, in _run_ninja_build
subprocess.run(
File "C:\Users\zhisong_liu\Miniconda3\envs\ngp\lib\subprocess.py", line 516, in run
raise CalledProcessError(retcode, process.args,
subprocess.CalledProcessError: Command '['ninja', '-v']' returned non-zero exit status 1.

I tried to fix the problem on the '['ninja', '-v']' by '['ninja', '--version']', but I got another problem as,

Traceback (most recent call last):
File "main_nerf.py", line 53, in
from nerf.network import NeRFNetwork
File "D:\torch-ngp\nerf\network.py", line 5, in
from encoding import get_encoder
File "D:\torch-ngp\encoding.py", line 5, in
from hashencoder import HashEncoder
File "D:\torch-ngp\hashencoder_init_.py", line 1, in
from .hashgrid import HashEncoder
File "D:\torch-ngp\hashencoder\hashgrid.py", line 9, in
from .backend import _backend
File "D:\torch-ngp\hashencoder\backend.py", line 6, in
_backend = load(name='_hash_encoder',
File "C:\Users\Miniconda3\envs\ngp\lib\site-packages\torch\utils\cpp_extension.py", line 1144, in load
return _jit_compile(
File "C:\Users\Miniconda3\envs\ngp\lib\site-packages\torch\utils\cpp_extension.py", line 1382, in _jit_compile
return _import_module_from_library(name, build_directory, is_python_module)
File "C:\Users\Miniconda3\envs\ngp\lib\site-packages\torch\utils\cpp_extension.py", line 1775, in _import_module_from_library
module = importlib.util.module_from_spec(spec)
ImportError: DLL load failed while importing _hash_encoder: The specified module could not be found.

What should I do?