NVIDIA Driver

[wyedoubleyou]

Hi, I wish to know if which version of Ubuntu and NVIDIA driver is compatible to run this project. Because I had problem on installing NVIDIA driver 470 in Ubuntu ver. 18.04 & running this project code in Ubuntu 20.04. Looking forward to the reply

[izakharkin]

Hello @yfung0, could you please provide the details of the problem you have while running the project?

[wyedoubleyou]

Hii @izakharkin, thank you so much for the reply.

I am using MSI GF63-Thin-11UD laptop, with GPU GeForce RTX 3050 Ti. I am using Ubuntu 20.04, installed Nvidia-Driver 470, CUDA Toolkit version 11.4, Cudnn version 8.2.2.

I had also completed the Setup, Build docker and Download data section.

When I executing the command python fit_outfit_code.py --config_name=outfit_code/psp in the docker container. It shows error as below:

(pbc) user@yw-GF63-Thin-11UD:/mounted/home/yw/Desktop/work/point_based_clothing$ python fit_outfit_code.py --config_name=outfit_code/psp

Traceback (most recent call last):
File "fit_outfit_code.py", line 32, in
from models.draping_network_wrapper import Wrapper
File "/mounted/home/yw/Desktop/work/point_based_clothing/src/models/draping_network_wrapper.py", line 10, in
from .draping_network import DrapingNetwork
File "/mounted/home/yw/Desktop/work/point_based_clothing/src/models/draping_network.py", line 8, in
from cloud_transformers.layers.multihead_ct_adain import MultiHeadUnionAdaIn, forward_style
File "/mounted/home/yw/Desktop/work/point_based_clothing/cloud_transformers/layers/multihead_ct_adain.py", line 4, in
from layers.cloud_transform import Splat, Slice, DifferentiablePositions
File "/mounted/home/yw/Desktop/work/point_based_clothing/cloud_transformers/layers/cloud_transform.py", line 7, in
from torch_scatter import scatter_max
File "/home/user/miniconda/envs/pbc/lib/python3.7/site-packages/torch_scatter/init.py", line 16, in
torch.ops.load_library(spec.origin)
File "/home/user/miniconda/envs/pbc/lib/python3.7/site-packages/torch/_ops.py", line 573, in load_library
ctypes.CDLL(path)
File "/home/user/miniconda/envs/pbc/lib/python3.7/ctypes/init.py", line 364, in init
self._handle = _dlopen(self._name, mode)
OSError: libcudart.so.10.2: cannot open shared object file: No such file or directory

Looking forward to your help.

[izakharkin]

Thank you for providing the details. Could you please also share the output of these commands inside the docker:

1. nvidia-smi
2. cat /usr/local/cuda/version.txt

[wyedoubleyou]

Here is the output :

(pbc) user@yw-GF63-Thin-11UD:/mounted/home/yw/Desktop/work/point_based_clothing$ nvidia-smi

Wed May 17 15:07:42 2023
+-----------------------------------------------------------------------------+
| NVIDIA-SMI 470.182.03 Driver Version: 470.182.03 CUDA Version: 11.4 |
|-------------------------------+----------------------+----------------------+
| GPU Name Persistence-M| Bus-Id Disp.A | Volatile Uncorr. ECC |
| Fan Temp Perf Pwr:Usage/Cap| Memory-Usage | GPU-Util Compute M. |
| | | MIG M. |
|===============================+======================+======================|
| 0 NVIDIA GeForce ... Off | 00000000:01:00.0 Off | N/A |
| N/A 40C P0 6W / N/A | 9MiB / 3913MiB | 0% Default |
| | | N/A |
+-------------------------------+----------------------+----------------------+

+-----------------------------------------------------------------------------+
| Processes: |
| GPU GI CI PID Type Process name GPU Memory |
| ID ID Usage |
|=============================================================================|
| 0 N/A N/A 1186 G /usr/lib/xorg/Xorg 4MiB |
| 0 N/A N/A 1973 G /usr/lib/xorg/Xorg 4MiB |
+-----------------------------------------------------------------------------+

(pbc) user@yw-GF63-Thin-11UD:/mounted/home/yw/Desktop/work/point_based_clothing$ cat /usr/local/cuda/version.txt
CUDA Version 10.1.243

[izakharkin]

Ok thank you, I will check it by myself

[arjundheek]

Is the issue solved?

[arjundheek]

Is this code really working?

[sahilshukla3003]

Hii @izakharkin, thank you so much for the reply.

I am using MSI GF63-Thin-11UD laptop, with GPU GeForce RTX 3050 Ti. I am using Ubuntu 20.04, installed Nvidia-Driver 470, CUDA Toolkit version 11.4, Cudnn version 8.2.2.

I had also completed the Setup, Build docker and Download data section.

When I executing the command python fit_outfit_code.py --config_name=outfit_code/psp in the docker container. It shows error as below:

(pbc) user@yw-GF63-Thin-11UD:/mounted/home/yw/Desktop/work/point_based_clothing$ python fit_outfit_code.py --config_name=outfit_code/psp

Traceback (most recent call last): File "fit_outfit_code.py", line 32, in from models.draping_network_wrapper import Wrapper File "/mounted/home/yw/Desktop/work/point_based_clothing/src/models/draping_network_wrapper.py", line 10, in from .draping_network import DrapingNetwork File "/mounted/home/yw/Desktop/work/point_based_clothing/src/models/draping_network.py", line 8, in from cloud_transformers.layers.multihead_ct_adain import MultiHeadUnionAdaIn, forward_style File "/mounted/home/yw/Desktop/work/point_based_clothing/cloud_transformers/layers/multihead_ct_adain.py", line 4, in from layers.cloud_transform import Splat, Slice, DifferentiablePositions File "/mounted/home/yw/Desktop/work/point_based_clothing/cloud_transformers/layers/cloud_transform.py", line 7, in from torch_scatter import scatter_max File "/home/user/miniconda/envs/pbc/lib/python3.7/site-packages/torch_scatter/init.py", line 16, in torch.ops.load_library(spec.origin) File "/home/user/miniconda/envs/pbc/lib/python3.7/site-packages/torch/_ops.py", line 573, in load_library ctypes.CDLL(path) File "/home/user/miniconda/envs/pbc/lib/python3.7/ctypes/init.py", line 364, in init self._handle = _dlopen(self._name, mode) OSError: libcudart.so.10.2: cannot open shared object file: No such file or directory

Looking forward to your help.

follow below steps

1. sudo find / -name "libcudart.so.11.0"
2. copy the path of libcudart.so.11.0 from your env
3. export LD_LIBRARY_PATH=/home/zzt-101/anaconda3/envs/cotton_env/lib:$LD_LIBRARY_PATH
4. put your path here instead of /home/zzt-101/anaconda3/envs/cotton_env/lib