実現したいこと
Tensorflow-gpu版でgpuを使用したい
困っていること
機械学習の学習速度が遅いと思い調べてみたところ、gpuを認識していないようでした。
機械学習の最初のアウトプット
python
1Using TensorFlow backend. 22019-01-28 14:14:17.288077: I tensorflow/core/platform/cpu_feature_guard.cc:141] Your CPU supports instructions that this TensorFlow binary was not compiled to use: AVX2 32019-01-28 14:14:18.122260: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1411] Found device 0 with properties: 4name: GeForce GT 630 major: 2 minor: 1 memoryClockRate(GHz): 1.62 5pciBusID: 0000:01:00.0 6totalMemory: 1.00GiB freeMemory: 822.90MiB 72019-01-28 14:14:18.156877: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1461] Ignoring visible gpu device (device: 0, name: GeForce GT 630, pci bus id: 0000:01:00.0, compute capability: 2.1) with Cuda compute capability 2.1. The minimum required Cuda capability is 3.7. 82019-01-28 14:14:18.201366: I tensorflow/core/common_runtime/gpu/gpu_device.cc:971] Device interconnect StreamExecutor with strength 1 edge matrix: 92019-01-28 14:14:18.202673: I tensorflow/core/common_runtime/gpu/gpu_device.cc:977] 0 102019-01-28 14:14:18.214962: I tensorflow/core/common_runtime/gpu/gpu_device.cc:990] 0: N
gpuが認識されているかチェックしたコード
Python
1from tensorflow.python.client import device_lib 2print(device_lib.list_local_devices())
console
12019-01-28 13:53:34.505727: I tensorflow/core/platform/cpu_feature_guard.cc:141] Your CPU supports instructions that this TensorFlow binary was not compiled to use: AVX2 22019-01-28 13:53:35.373934: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1411] Found device 0 with properties: 3name: GeForce GT 630 major: 2 minor: 1 memoryClockRate(GHz): 1.62 4pciBusID: 0000:01:00.0 5totalMemory: 1.00GiB freeMemory: 822.90MiB 62019-01-28 13:53:35.374676: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1461] Ignoring visible gpu device (device: 0, name: GeForce GT 630, pci bus id: 0000:01:00.0, compute capability: 2.1) with Cuda compute capability 2.1. The minimum required Cuda capability is 3.7. 72019-01-28 13:53:35.390906: I tensorflow/core/common_runtime/gpu/gpu_device.cc:971] Device interconnect StreamExecutor with strength 1 edge matrix: 82019-01-28 13:53:35.392599: I tensorflow/core/common_runtime/gpu/gpu_device.cc:977] 0 92019-01-28 13:53:35.394655: I tensorflow/core/common_runtime/gpu/gpu_device.cc:990] 0: N 10[name: "/device:CPU:0" 11device_type: "CPU" 12memory_limit: 268435456 13locality { 14} 15incarnation: 14268899556145551426 16]
やったこと
Tensorflowインストール時には対応したCUDAとcuDNNをインストールしました。
環境
GPU:GeForce GT630
CUDA 9.0
nvccv
1nvcc: NVIDIA (R) Cuda compiler driver 2Copyright (c) 2005-2017 NVIDIA Corporation 3Built on Fri_Sep__1_21:08:32_Central_Daylight_Time_2017 4Cuda compilation tools, release 9.0, V9.0.176
cuDNN 7.0.5
pip
pip
1 2Package Version 3------------------- ---------- 4absl-py 0.6.0 5astor 0.7.1 6certifi 2018.10.15 7chardet 3.0.4 8cloudpickle 0.6.1 9cycler 0.10.0 10dask 0.20.1 11decorator 4.3.0 12flickrapi 2.4.0 13gast 0.2.0 14grpcio 1.15.0 15h5py 2.8.0 16idna 2.7 17Keras 2.2.4 18Keras-Applications 1.0.6 19Keras-Preprocessing 1.0.5 20kiwisolver 1.0.1 21Markdown 3.0.1 22matplotlib 3.0.2 23networkx 2.2 24numpy 1.15.3 25oauthlib 2.1.0 26opencv-python 3.4.4.19 27pandas 0.23.4 28Pillow 5.3.0 29pip 18.1 30protobuf 3.6.1 31py2exe 0.9.2.2 32pyparsing 2.2.2 33python-dateutil 2.7.3 34pytz 2018.7 35PyWavelets 1.0.1 36pywin32 224 37PyYAML 3.13 38requests 2.20.1 39requests-oauthlib 1.0.0 40requests-toolbelt 0.8.0 41scikit-image 0.14.1 42scikit-learn 0.20.0 43scipy 1.1.0 44setuptools 39.1.0 45six 1.11.0 46tensorboard 1.11.0 47tensorflow-gpu 1.11.0 48termcolor 1.1.0 49toolz 0.9.0 50urllib3 1.24.1 51Werkzeug 0.14.1 52wheel 0.32.2
教えていただけますでしょうか。
回答1件
あなたの回答
tips
プレビュー