CUDA | Hough Transform Tracking using Thrust

I have a problem about not accessing GPU in PyCharm and I use NVIDIA as GPU.

I installed tensorflow-gpu in Python Interpreter of Setting part in Pycharm and then I run the code but I still cannot access it.

I wonder if I should use CUDA library? How can I fix it?

Here is my code snippet which is shown below.

Machine Setting:

• GPU: GeForce RTX 3060

• Driver Version: 460.73.01

• CUDA Driver Veresion: 11.2

• Tensorflow: tensorflow-gpu 1.14.0

• CUDA Runtime Version: 10.0

• cudnn: 7.4.1

Note:

1. CUDA Runtime and cudnn version fits the guide from Tensorflow official documentation.
2. I've also tried for TensorFlow-gpu = 2.0, still the same problem.

Problem:

I am using Tensorflow for an objection detection task. My situation is that the program will stuck at

2021-06-05 12:16:54.099778: I tensorflow/stream_executor/platform/default/dso_loader.cc:42] Successfully opened dynamic library libcublas.so.10

for several minutes.

And then stuck at next loading process

2021-06-05 12:21:22.212818: I tensorflow/stream_executor/platform/default/dso_loader.cc:42] Successfully opened dynamic library libcudnn.so.7

for even longer time. You may check log.txt for log details.

After waiting for around 30 mins, the program will start to running and WORK WELL.

However, whenever program invoke self.session.run(...), it will load the same two library related to cuda (libcublas and libcudnn) again, which is time-wasted and annoying.

I am confused that where the problem comes from and how to resolve it. Anyone could help?

Discussion Issue on Github

===================================

Update

After @talonmies 's help, the problem was resolved by resetting the environment with correct version matching among GPU, CUDA, cudnn and tensorflow. Now it works smoothly.

I am programming in Python 3.8 with Tensorflow installed along with my natural language processing project. When I want to begin the training phase, I get this message right before I begin...

I am trying to use thrust with Opencv classes. The final code will be more complicated including using device memory but this simple example does not build successfully.

I'm trying to create a pseudo-random generator API, but numbers generated by xorshift have unrandom nature. You can see the algorithm and tests here:

I have a pyTorch-code to train a model that should be able to detect placeholder-images among product-images. I didn't write the code by myself as I am very unexperienced with CNNs and Machine Learning.

My boss told me to calculate the f1-score for that model and i found out that the formula for that is ((precision * recall)/(precision + recall)) but I don't know how I get precision and recall. Is someone able to tell me how I can get those two parameters from that following code? (Sorry for the long piece of code, but I didn't really know what is necessary and what isn't)

I want to force the Huggingface transformer (BERT) to make use of CUDA. nvidia-smi showed that all my CPU cores were maxed out during the code execution, but my GPU was at 0% utilization. Unfortunately, I'm new to the Hugginface library as well as PyTorch and don't know where to place the CUDA attributes device = cuda:0 or .to(cuda:0).

The code below is basically a customized part from german sentiment BERT working example

I'm new on PyTorch and I'm trying to code with it so I have a function called OH which tack a number and return a vector like this

My task is to calculate the approximate value of pi with an accuracy of at least 10^-6. The Monte Carlo algorithm does not provide the required accuracy. I need to use the calculation only through the volume of the sphere. What do you advise? I would be glad to see examples of code in CUDA or pure C++. Thank you.