Tensorflow停止识别我的GPU（以前可以工作）。
代码：
import os
import tensorflow as tf
os.environ["CUDA_VISIBLE_DEVICES"] = "0"
a = tf.constant([1.0, 2.0, 3.0, 4.0, 5.0, 6.0], shape=[2, 3], name='a')
b = tf.constant([1.0, 2.0, 3.0, 4.0, 5.0, 6.0], shape=[3, 2], name='b')
c = tf.matmul(a, b)
print(c)
输出：
2019-05-24 10:16:30.726799: I tensorflow/core/platform/cpu_feature_guard.cc:142] Your CPU supports instructions that this TensorFlow binary was not compiled to use: AVX2 FMA
2019-05-24 10:16:30.740369: I tensorflow/stream_executor/platform/default/dso_loader.cc:42] Successfully opened dynamic library libcuda.so.1
2019-05-24 10:16:30.777178: E tensorflow/stream_executor/cuda/cuda_driver.cc:318] failed call to cuInit: CUDA_ERROR_INVALID_DEVICE: invalid device ordinal
2019-05-24 10:16:30.777232: I tensorflow/stream_executor/cuda/cuda_diagnostics.cc:166] retrieving CUDA diagnostic information for host: joao-Yoga730
2019-05-24 10:16:30.777242: I tensorflow/stream_executor/cuda/cuda_diagnostics.cc:173] hostname: joao-Yoga730
2019-05-24 10:16:30.777331: I tensorflow/stream_executor/cuda/cuda_diagnostics.cc:197] libcuda reported version is: 410.48.0
2019-05-24 10:16:30.777364: I tensorflow/stream_executor/cuda/cuda_diagnostics.cc:201] kernel reported version is: 410.48.0
2019-05-24 10:16:30.777371: I tensorflow/stream_executor/cuda/cuda_diagnostics.cc:308] kernel version seems to match DSO: 410.48.0
2019-05-24 10:16:30.799754: I tensorflow/core/platform/profile_utils/cpu_utils.cc:94] CPU Frequency: 1992000000 Hz
2019-05-24 10:16:30.801312: I tensorflow/compiler/xla/service/service.cc:162] XLA service 0x55e937b11800 executing computations on platform Host. Devices:
2019-05-24 10:16:30.801384: I tensorflow/compiler/xla/service/service.cc:169] StreamExecutor device (0): <undefined>, <undefined>
tf.Tensor(
[[22. 28.]
[49. 64.]], shape=(2, 2), dtype=float32)
Process finished with exit code 0
nvidia-smi输出：
+-----------------------------------------------------------------------------+
| NVIDIA-SMI 410.48 Driver Version: 410.48 |
|-------------------------------+----------------------+----------------------+
| GPU Name Persistence-M| Bus-Id Disp.A | Volatile Uncorr. ECC |
| Fan Temp Perf Pwr:Usage/Cap| Memory-Usage | GPU-Util Compute M. |
|===============================+======================+======================|
| 0 GeForce GTX 1050 Off | 00000000:3B:00.0 Off | N/A |
| N/A 52C P3 N/A / N/A | 500MiB / 4042MiB | 4% Default |
+-------------------------------+----------------------+----------------------+
+-----------------------------------------------------------------------------+
| Processes: GPU Memory |
| GPU PID Type Process name Usage |
|=============================================================================|
| 0 1474 G /usr/lib/xorg/Xorg 296MiB |
| 0 1650 G /usr/bin/gnome-shell 145MiB |
| 0 2107 G ...uest-channel-token=16113907172495644879 55MiB |
| 0 4229 G /snap/pycharm-community/128/jre64/bin/java 2MiB |
+-----------------------------------------------------------------------------+
Ubuntu 18.04 Tensorflow'2.0.0-alpha0'
有什么想法吗？
p.s。我尝试做sudo apt install nvidia-modprobe。没用。