Ubuntu 當初同時按照了 cuda9.0 和 cuda10.0，但因爲是安裝的是雙系統windows10 + Ubuntu16.04,在C盤分割了 40G個給Ubuntu，安裝2個cuda太過佔存儲空間,/dev/sda6 use%已經到90%多，需要刪除一個多餘的cuda10.0.
wu@wu-X555LF:~$ cd /usr/local/cuda/bin
wu@wu-X555LF:/usr/local/cuda-10.0/bin$ sudo uninstall_cuda_10.0.pl
sudo: uninstall_cuda_10.0.pl: command not found
在/usr/local/cuda/bin 目錄下,有cuda 自帶的卸載工具uninstall_cuda_10.0.pl
-
cd /usr/local/cuda/bin -
sudo ./uninstall_cuda_10.0.pl
刪除完cuda10.0，通過df -h查掛載情況：
wu@wu-X555LF:/usr/local/cuda-10.0/bin$ df -h
Filesystem Size Used Avail Use% Mounted on
udev 5.8G 0 5.8G 0% /dev
tmpfs 1.2G 9.4M 1.2G 1% /run
/dev/sda6 19G 14G 4.3G 76% /
tmpfs 5.9G 37M 5.8G 1% /dev/shm
tmpfs 5.0M 4.0K 5.0M 1% /run/lock
tmpfs 5.9G 0 5.9G 0% /sys/fs/cgroup
/dev/sda8 9.1G 5.8G 2.9G 68% /home
/dev/sdb1 256M 31M 226M 12% /boot/efi
tmpfs 1.2G 80K 1.2G 1% /run/user/1000
/dev/sda2 100G 77G 23G 78% /media/wu/22FC52F5FC52C329
執行情況，Your CPU supports instructions that this TensorFlow binary was not compiled to use: AVX2 FMA，還是出現問題：
wu@wu-X555LF:~/YOLOv3_TensorFlow-master$ python test_single_image.py ./data/demo_data/messi.jpg
2019-07-20 22:55:38.464088: I tensorflow/core/platform/cpu_feature_guard.cc:140] Your CPU supports instructions that this TensorFlow binary was not compiled to use: AVX2 FMA
2019-07-20 22:55:38.587955: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:898] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero
2019-07-20 22:55:38.588425: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1212] Found device 0 with properties:
name: GeForce 930M major: 5 minor: 0 memoryClockRate(GHz): 0.941
pciBusID: 0000:04:00.0
totalMemory: 1.96GiB freeMemory: 1.82GiB
2019-07-20 22:55:38.588448: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1312] Adding visible gpu devices: 0
2019-07-20 22:55:40.336327: I tensorflow/core/common_runtime/gpu/gpu_device.cc:993] Creating TensorFlow device (/job:localhost/replica:0/task:0/device:GPU:0 with 1588 MB memory) -> physical GPU (device: 0, name: GeForce 930M, pci bus id: 0000:04:00.0, compute capability: 5.0)
2019-07-20 22:55:50.530431: E tensorflow/stream_executor/cuda/cuda_dnn.cc:378] Loaded runtime CuDNN library: 7500 (compatibility version 7500) but source was compiled with 7004 (compatibility version 7000). If using a binary install, upgrade your CuDNN library to match. If building from sources, make sure the library loaded at runtime matches a compatible version specified during compile configuration.
2019-07-20 22:55:50.531361: F tensorflow/core/kernels/conv_ops.cc:717] Check failed: stream->parent()->GetConvolveAlgorithms( conv_parameters.ShouldIncludeWinogradNonfusedAlgo<T>(), &algorithms)