Я пытаюсь запустить Tensorflow на своем графическом процессоре в Ubuntu 20.04. Вот что я сделал до сих пор:
- Установлен драйвер nvidia 460
- Установлены cuda toolkit 10.1 и cuDNN v7.6.5.
- Установил Tensorflow с помощью следующей команды: pip3 install tensorflow == 2.3.0
Когда я запускаю следующий код, чтобы проверить, находит ли TF графический процессор:
import tensorflow as tf
sess = tf.compat.v1.Session(config=tf.compat.v1.ConfigProto(log_device_placement=True))
Я получаю сообщение об ошибке:
2021-03-27 10:16:56.375617: I tensorflow/stream_executor/platform/default/dso_loader.cc:48] Successfully opened dynamic library libcudart.so.10.1
2021-03-27 10:16:57.119060: I tensorflow/core/platform/cpu_feature_guard.cc:142] 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.
2021-03-27 10:16:57.141600: I tensorflow/core/platform/profile_utils/cpu_utils.cc:104] CPU Frequency: 2899885000 Hz
2021-03-27 10:16:57.141871: I tensorflow/compiler/xla/service/service.cc:168] XLA service 0x443c730 initialized for platform Host (this does not guarantee that XLA will be used). Devices:
2021-03-27 10:16:57.141906: I tensorflow/compiler/xla/service/service.cc:176] StreamExecutor device (0): Host, Default Version
2021-03-27 10:16:57.143821: I tensorflow/stream_executor/platform/default/dso_loader.cc:48] Successfully opened dynamic library libcuda.so.1
2021-03-27 10:16:57.194105: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:982] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero
2021-03-27 10:16:57.194532: I tensorflow/compiler/xla/service/service.cc:168] XLA service 0x44d6e60 initialized for platform CUDA (this does not guarantee that XLA will be used). Devices:
2021-03-27 10:16:57.194550: I tensorflow/compiler/xla/service/service.cc:176] StreamExecutor device (0): GeForce 940MX, Compute Capability 5.0
2021-03-27 10:16:57.194686: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:982] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero
2021-03-27 10:16:57.195008: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1716] Found device 0 with properties:
pciBusID: 0000:01:00.0 name: GeForce 940MX computeCapability: 5.0
coreClock: 1.189GHz coreCount: 3 deviceMemorySize: 3.95GiB deviceMemoryBandwidth: 29.80GiB/s
2021-03-27 10:16:57.195033: I tensorflow/stream_executor/platform/default/dso_loader.cc:48] Successfully opened dynamic library libcudart.so.10.1
2021-03-27 10:16:57.285500: I tensorflow/stream_executor/platform/default/dso_loader.cc:48] Successfully opened dynamic library libcublas.so.10
2021-03-27 10:16:57.289676: I tensorflow/stream_executor/platform/default/dso_loader.cc:48] Successfully opened dynamic library libcufft.so.10
2021-03-27 10:16:57.290460: I tensorflow/stream_executor/platform/default/dso_loader.cc:48] Successfully opened dynamic library libcurand.so.10
2021-03-27 10:16:57.295273: I tensorflow/stream_executor/platform/default/dso_loader.cc:48] Successfully opened dynamic library libcusolver.so.10
2021-03-27 10:16:57.298153: I tensorflow/stream_executor/platform/default/dso_loader.cc:48] Successfully opened dynamic library libcusparse.so.10
Bus error (core dumped)
Как я могу это исправить?