386 questions
0
votes
0
answers
42
views
Google Colab AlexNet.ipynb: OSError: libcudnn.so.8: No such file or directory
I'm trying to learn and implement AlexNet and I'm using this Google Colab notebook to help: https://colab.research.google.com/github/d2l-ai/d2l-en-colab/blob/master/chapter_convolutional-modern/...
-4
votes
1
answer
123
views
Failing to Finalize Execution Plan Using cuDNN Backend to Create a Fused Attention fprop Graph
I am working on implementing the Fused Attention fprop graph pattern. As of now I am only combining two matrix multiplications, meaning g3 and g4 are empty. I believe I have also matched all the ...
- 63
2
votes
2
answers
3k
views
jax library error jaxlib.xla_extension.XlaRuntimeError: FAILED_PRECONDITION: DNN library initialization failed
I am working on a code that uses the Jax library, and I run into this error over and over again no matter how I tried to configure my environment:
2024-08-20 16:26:58.037892: E external/xla/xla/...
- 57
0
votes
1
answer
496
views
TensorFlow 2.16.2 does not detecting my GPU
I got RTX 4060 with driver 560.76.
I installed CUDA 11.8 and copied cuDNN 8.9.7. files to the correct directories in the CUDA installation folder.
I added the right paths to the System variables ...
- 73
0
votes
0
answers
82
views
Error creating engineConfig with cuDNN Graph API in custom implementation
I'm working on a custom implementation using cuDNN's Graph API, but I'm encountering an error when creating the engineConfig. I'm trying to implement this directly using cuDNN's backend API calls, ...
- 1,357
0
votes
0
answers
50
views
Unable to Run TensorFlow in GPU Mode Despite Correct Setup
ISSUE:
I have been trying without success to run TensorFlow in GPU mode. When I run the following code to check for available devices:
from tensorflow.python.client import device_lib
def ...
user11694357
1
vote
1
answer
412
views
Replacement for deprecated cuDNN functions - official NVIDIA migration path?
I have successfully tested a small C++ program which uses the NVIDIA cuDNN API to run code on a GPU. The cuDNN API level is 9.1.1. The program compiles, links and runs OK.
What puzzles me is that ...
0
votes
0
answers
210
views
TensorFlow ImportError: undefined symbol: _ZTIN6snappy4SinkE
I'm trying to import TensorFlow in my conda Python environment, but I'm encountering the following error:
---------------------------------------------------------------------------
ImportError ...
-3
votes
1
answer
617
views
dlib fails to install with cuda - says I am missing cudnn
Windows 11
Visual Studio 17 2022
cuDNN v9.0
Cuda v12.3
I have set the enviromental variables to:
$env:CUDA_PATH = "C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v12.3"
Additionally, my $...
- 81
0
votes
0
answers
1k
views
CUDA_ERROR_INVALID_SOURCE: device kernel image is invalid
I'm trying using cupy in my docker container.
I use to containers which one is for CUDA and cuDNN, and the other is for cupy.
I tried this code.
import cupy as cp
cupy_array = cp.array([1, 2, 3])
...
- 13
0
votes
0
answers
2k
views
CUDA compatibility with Tensorflow and CuDNN for Deep Learning Models
I tried several combinations from this list to try and activate GPU acceleration for deep learning. However, nothing seems to work with tensorflow.
The CUDA installation works fine (by fine I mean ...
0
votes
2
answers
1k
views
I have both cudatoolkit and cudnn in my conda environment, but pytorch is still using my CPU
(my_project_env) C:\\Users...\\pythonProject3\>conda list cudatoolkit
# packages in environment at C:\\ProgramData\\anaconda3\\envs\\my_project_env:
#
# Name Version ...
2
votes
1
answer
5k
views
Loaded runtime CuDNN library: 8.5.0 but source was compiled with: 8.6.0
I have a problem with tensorflow. For weeks everything was working and now I get the error:
Loaded runtime CuDNN library: 8.5.0 but source was compiled with: 8.6.0. CuDNN library needs to have ...
- 81
-1
votes
1
answer
274
views
Dlerror: cudnn64_7.dll not found
When I run tf.test.is_gpu_availabe() it returns false and the only problem that it is showing that cudnn64_7.dll not found
I am using python 3.8.17 for tensorflow 2.3.0
Cuda-10.1, CudNN-8.0.5
GPU-...
0
votes
1
answer
425
views
How can i update cuDNN on azure ML studio from 8.2 to 8.6 while running on pytorch and tensorflow kernel
I have uninstalled and installed cuDNN but still it is giving error. I have tried to downgrade tensorflow and pytorch but azure runs on the default version.This is the error am facing
I have installed ...
0
votes
0
answers
119
views
RNN Keras Error || No OpKernel was registered to support Op 'CudnnRNN' used by {{node CudnnRNN}} with these attrs
I am currently working on a RNN using Keras, but I am stuck with this error:
---------------------------------------------------------------------------
ValueError ...
-1
votes
1
answer
422
views
Run Pytorch examples with Pytorch build from source
I have build pytorch 2.0.1 from source. Using cuda 11.7, cudnn v8, and the driver for the nvidia GPU is 515.43.04 (CUDA version 11.7). Altough Pytorch seems to build successfully when I am trying to ...
- 41
2
votes
2
answers
1k
views
Azure VM Loaded runtime CuDNN library: 8.2.4 but source was compiled with: 8.6.0
I have tried to fit a Keras model on a notebook in Microsoft Azure Machine Learning Studio GPU machine. I have received an error similar to what was described here:
2023-04-27 09:56:21.098249: E ...
- 763
3
votes
0
answers
203
views
Why does the CuDNN LSTM requires two biases b_ih and b_hh
The question was asked quite often for the higher-level frameworks:
PyTorch: Why RNN need two bias vectors?
PyTorch: Why RNN needs two biases?
TF/Keras: Why CuDNNLSTM has more parameters than LSTM?
...
- 68.9k
1
vote
0
answers
712
views
Why is CUDA with pytorch freezing and work worse than cpu on my device? Is this a configuration issue or a hardware issue?
I am trying to retrain the last layer of ResNet18 but running into problems using CUDA. I am not hearing the GPU and in Task Manager GPU usage is minimal when running with CUDA. I increased the ...
- 11
-1
votes
1
answer
807
views
OpenPose using Ubuntu 20 CuDNN 8 CMAKE can't find CuDNN
I have CuDNN 8 installed on Ubuntu 20.04and I want to install OpenPose But I get a message the CuDNN was not Found:
~/Downloads/openpose/build((HEAD detached at v1.7.0))$ cmake .. -DBUILD_PYTHON=ON -...
- 4,125
11
votes
1
answer
22k
views
No matching distribution found in the installation of the cuDNN for TensorFlow v2.12 in Anaconda
I have tried to install TensorFlow v2.12 in Anaconda with Python 3.9.16 and Windows 10 OS with pip v23.0.1. I need Tensforflow v2.12 for my GPU (RTX4080), and only this version works with my GPU, ...
-2
votes
1
answer
7k
views
On windows11, nvcc cannot show the correct version of CUDA
The version I originally installed is 11.6. To match the tensorflow2.11.0, I had to install the v11.2 of CUDA, during which I first uinstall the newer version of CUDA(every thing about it) and then ...
- 51
1
vote
0
answers
49
views
Build cuDNN framework with CPU fallback
I am trying to implement some neural networks for training and inference using C++. It should work on GPU (if available, with cuDNN) and CPU (if GPU is not available).
All modern frameworks support ...
- 343
1
vote
1
answer
5k
views
Conda cannot see Cuda version
I am trying to install the newest Tensorflow GPU version to an Ubuntu environment. The Cuda drivers are correctly installed and working, which I can confirm with the following commands:
nvcc --version
...
- 3,679