# TensorFlow Install error **URL:** \ **Category:** General Discussion\ **Tags:** gpu, ubuntu, install\ **Created:** [October 30, 2023, 3:43am UTC](https://discuss.ai.google.dev/t/tensorflow-install-error/31117 "2023-10-30T03:43:11Z")\ **Posts on this page:** 5\ **Page:** 1
**Author:** ![min\_guell](https://d1dlmcr85iqnpo.cloudfront.net/user_avatar/discuss.ai.google.dev/min_guell/32/17849_2.png) [@min\_guell](https://discuss.ai.google.dev/u/min_guell)\ **Post date:** [October 30, 2023, 3:43am UTC](https://discuss.ai.google.dev/t/tensorflow-install-error/31117/1 "2023-10-30T03:43:11Z")
Hey, Im trying to use Tensorflow 2.14.0 in my Ubuntu 22.04 but he doesnt recognize my gpu (NVIDIA 4070ti). My cuda version is 11.5, python 3.10.12, pip 22.0.2 and TensorRT 8.5.3.1. When I try to run, receive this: E tensorflow/compiler/xla/stream\_executor/cuda/cuda\_dnn.cc:9342] Unable to register cuDNN factory: Attempting to register factory for plugin cuDNN when one has already been registered E tensorflow/compiler/xla/stream\_executor/cuda/cuda\_fft.cc:609] Unable to register cuFFT factory: Attempting to register factory for plugin cuFFT when one has already been registered E tensorflow/compiler/xla/stream\_executor/cuda/cuda\_blas.cc:1518] Unable to register cuBLAS factory: Attempting to register factory for plugin cuBLAS when one has already been registered I tensorflow/core/platform/cpu\_feature\_guard.cc:182] This TensorFlow binary is optimized to use available CPU instructions in performance-critical operations. To enable the following instructions: AVX2 FMA, in other operations, rebuild TensorFlow with the appropriate compiler flags. W tensorflow/compiler/tf2tensorrt/utils/py\_utils.cc:38] TF-TRT Warning: Could not find TensorRT I tensorflow/compiler/xla/stream\_executor/cuda/cuda\_gpu\_executor.cc:894] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero. See more at [linux/Documentation/ABI/testing/sysfs-bus-pci at v6.0 · torvalds/linux · GitHub](https://github.com/torvalds/linux/blob/v6.0/Documentation/ABI/testing/sysfs-bus-pci#L344-L355) W tensorflow/core/common\_runtime/gpu/gpu\_device.cc:2211] Cannot dlopen some GPU libraries. Please make sure the missing libraries mentioned above are installed properly if you would like to use GPU. Follow the guide at [https://www.tensorflow.org/install/gpu](https://www.tensorflow.org/install/gpu) for how to download and setup the required libraries for your platform. Skipping registering GPU devices… Also, when I try run this: # verify available gpu physical\_devices = tf.config.list\_physical\_devices(‘GPU’) if len(physical\_devices) == 0: print(“Error”) else: # Tensorflow config to use gpu tf.config.experimental.set\_memory\_growth(physical\_devices[0], True) He prints the error message ---
**Author:** ![Renu\_Patel](https://d1dlmcr85iqnpo.cloudfront.net/user_avatar/discuss.ai.google.dev/renu_patel/32/8957_2.png) [@Renu\_Patel](https://discuss.ai.google.dev/u/Renu_Patel)\ **Post date:** [October 30, 2023, 3:54pm UTC](https://discuss.ai.google.dev/t/tensorflow-install-error/31117/3 "2023-10-30T15:54:37Z")
Hi @min_guell Welcome to the TensorFlow Forum! The installed CUDA version is not compatible with Tensorflow 2.14. You may need to install `CUDA 11.8` and `cuDNN 8.7` for `Tensorflow 2.14` as mentioned in this [tested build configuration](https://www.tensorflow.org/install/source#linux). Please refer to this [TF install](https://www.tensorflow.org/install/pip) page and check all the **Hardware/Software requirements** for GPU support along with following the **step step by instructions** mentioned to install Tensorflow with GPU. Please let us know if the issue still persists. Thank you. ---
**Author:** ![Alvin\_Liu](https://d1dlmcr85iqnpo.cloudfront.net/user_avatar/discuss.ai.google.dev/alvin_liu/32/7421_2.png) [@Alvin\_Liu](https://discuss.ai.google.dev/u/Alvin_Liu)\ **Post date:** [November 22, 2023, 4:31am UTC](https://discuss.ai.google.dev/t/tensorflow-install-error/31117/5 "2023-11-22T04:31:13Z")
I think I have the same problem. GPU: A800 DriveVersion: 525.105.17 CUDA Version: 12.0 I’ve installed such versions below `tensorflow 2.9.0 / 2.12/ 2.13.1 / 2.14.0 ` But it seems like tensorflow could not detect GPU because: `available = tf.test.is_gpu_available()` `is_cuda_gpu_available = tf.test.is_gpu_available(cuda_only=True)` `is_cuda_gpu_min_3 = tf.test.is_gpu_available(True, (3,0))` All were “False”. When I tried `pip install tensorflow[and-cuda]==2.15.0` It exactly returns “True” But still received the error message: ```auto E external/local_xla/xla/stream_executor/cuda/cuda_dnn.cc:9261] Unable to register cuDNN factory: Attempting to register factory for plugin cuDNN when one has already been registered E external/local_xla/xla/stream_executor/cuda/cuda_fft.cc:607] Unable to register cuFFT factory: Attempting to register factory for plugin cuFFT when one has already been registered E external/local_xla/xla/stream_executor/cuda/cuda_blas.cc:1515] Unable to register cuBLAS factory: Attempting to register factory for plugin cuBLAS when one has already been registered ``` ---
**Author:** ![adiazmazzola](https://d1dlmcr85iqnpo.cloudfront.net/user_avatar/discuss.ai.google.dev/adiazmazzola/32/14133_2.png) [@adiazmazzola](https://discuss.ai.google.dev/u/adiazmazzola)\ **Post date:** [December 26, 2023, 1:16am UTC](https://discuss.ai.google.dev/t/tensorflow-install-error/31117/6 "2023-12-26T01:16:40Z")
Hello I’ve got the same problem and solved (not yet the -1 NUMA nodes) following this [YouTube](https://www.youtube.com/watch?v=4p4b9qT5S6w&t=464s) tutorial, First I wiped out cuda tool kit on my system with: sudo apt-get --purge remove “_cuda_” “_cublas_” “_cufft_” “_cufile_” “_curand_” “_cusolver_” “_cusparse_” “_gds-tools_” “_npp_” “_nvjpeg_” “nsight\*” “_nvvm_” and then follow then follow the cudnn installation guide but directing to the specific folder of the new cuda 11.8 installed on my system, like this: $ sudo cp cudnn-_-archive/include/cudnn_.h /usr/local/cuda-11.8/include $ sudo cp -P cudnn-_-archive/lib/libcudnn_ /usr/local/cuda-11.8/lib64 $ sudo chmod a+r /usr/local/cuda/include/cudnn\*.h /usr/local/cuda-11.8/lib64/libcudnn\* By the way, the for venv I used conda: $ conda create -n tf python=3.10 $ conda activate tf $ pip install --upgrade pip $ pip install -U setuptools wheel $ pip install tensorflow==2.12.0 if you find any solution for NUMA nodes please share it 🙂 Have a nice one! ---
**Author:** ![Srikanth\_NC](https://d1dlmcr85iqnpo.cloudfront.net/user_avatar/discuss.ai.google.dev/srikanth_nc/32/14365_2.png) [@Srikanth\_NC](https://discuss.ai.google.dev/u/Srikanth_NC)\ **Post date:** [January 31, 2024, 6:42am UTC](https://discuss.ai.google.dev/t/tensorflow-install-error/31117/8 "2024-01-31T06:42:02Z")
Hi I have installed tensorflow-gpu - 2.10.1 and I have cuda version 11.4 When I try to import tensorflow I get the following error. Can anyone please guide me to solve the error. But suprisingly i can see the number of physical gpu using tensorflow command. I tensorflow/core/platform/cpu\_feature\_guard.cc:193] This TensorFlow binary is optimized with oneAPI Deep Neural Network Library (oneDNN) to use the following CPU instructions in performance-critical operations: AVX2 FMA To enable them in other operations, rebuild TensorFlow with the appropriate compiler flags. E tensorflow/stream\_executor/cuda/cuda\_blas.cc:2981] Unable to register cuBLAS factory: Attempting to register factory for plugin cuBLAS when one has already been registered W tensorflow/stream\_executor/platform/default/dso\_loader.cc:64] Could not load dynamic library ‘libnvinfer.so.7’; dlerror: libnvinfer.so.7: cannot open shared object file: No such file or directory W tensorflow/stream\_executor/platform/default/dso\_loader.cc:64] Could not load dynamic library ‘libnvinfer\_plugin.so.7’; dlerror: libnvinfer\_plugin.so.7: cannot open shared object file: No such file or directory W tensorflow/compiler/tf2tensorrt/utils/py\_utils.cc:38] TF-TRT Warning: Cannot dlopen some TensorRT libraries. If you would like to use Nvidia GPU with TensorRT, please make sure the missing libraries mentioned above are installed properly.