BIGBALLON

• [Tensorflow中使用指定的GPU及GPU显存][1]

//终端执行程序时设置使用的GPU
CUDA_VISIBLE_DEVICES=1 python my_script.py

//python代码中设置使用的GPU
import os
os.environ["CUDA_VISIBLE_DEVICES"] = "2"

BIGBALLON

import keras
import numpy as np
from keras.datasets import cifar10
from keras.preprocessing.image import ImageDataGenerator
from keras.layers.normalization import BatchNormalization
from keras.layers import Conv2D, Dense, Input, add, Activation, GlobalAveragePooling2D
from keras.callbacks import LearningRateScheduler, TensorBoard, ModelCheckpoint
from keras.models import Model
from keras import optimizers, regularizers
from keras import backend as K

BIGBALLON

from tensorflow.examples.tutorials.mnist import input_data
import tensorflow as tf

def main(_):
  mnist = input_data.read_data_sets("./data", one_hot=True)
  x = tf.placeholder(tf.float32, [None, 784])
  W = tf.Variable(tf.truncated_normal(shape=[784, 10], stddev=0.1))
  b = tf.Variable(tf.constant(0.1, shape=[10]))
  y = tf.matmul(x, W) + b
  y_ = tf.placeholder(tf.float32, [None, 10])

BIGBALLON

(dp) dl2017@mtk:~/Desktop/horovod/examples$ strace -f -e 'trace=!poll'     mpirun -np 2     -H 192.168.2.243:1,192.168.3.246:1     -bind-to none -map-by slot     -x NCCL_DEBUG=INFO -x LD_LIBRARY_PATH     python3 keras_mnist_advanced.pyexecve("/usr/local/bin/mpirun", ["mpirun", "-np", "2", "-H", "192.168.2.243:1,192.168.3.246:1", "-bind-to", "none", "-map-by", "slot", "-x", "NCCL_DEBUG=INFO", "-x", "LD_LIBRARY_PATH", "python3", "keras_mnist_advanced.py"], [/* 36 vars */]) = 0
brk(NULL)                               = 0x10ec000
access("/etc/ld.so.nohwcap", F_OK)      = -1 ENOENT (No such file or directory)
mmap(NULL, 8192, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_ANONYMOUS, -1, 0) = 0x7f32b0cc6000
access("/etc/ld.so.preload", R_OK)      = -1 ENOENT (No such file or directory)
open("/usr/local/cuda-8.0/lib64/tls/x86_64/libopen-rte.so.40", O_RDONLY|O_CLOEXEC) = -1 ENOENT (No such file or directory)
stat("/usr/local/cuda-8.0/lib64/tls/x86_64", 0x7ffdaecc9af0) = -1 ENOENT (No such file or directory)
open("/usr/local/c

BIGBALLON

(dp) dl2017@mtk:~/Desktop/horovod/examples$ strace mpirun --prefix /usr/local \
>     -np 2 \
>     -H 192.168.2.243:1,192.168.3.246:1 \
>     -bind-to none -map-by slot \
>     -x NCCL_DEBUG=INFO -x LD_LIBRARY_PATH \
>     python3 keras_mnist_advanced.py
execve("/usr/local/bin/mpirun", ["mpirun", "--prefix", "/usr/local", "-np", "2", "-H", "192.168.2.243:1,192.168.3.246:1", "-bind-to", "none", "-map-by", "slot", "-x", "NCCL_DEBUG=INFO", "-x", "LD_LIBRARY_PATH", "python3", ...], [/* 36 vars */]) = 0
brk(NULL)                               = 0x19c7000
access("/etc/ld.so.nohwcap", F_OK)      = -1 ENOENT (No such file or directory)
mmap(NULL, 8192, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_ANONYMOUS, -1, 0) = 0x7f0c9c190000

BIGBALLON

dl2017@mtk:~$ strace mpirun -np 2     -H 192.168.2.243:1,192.168.3.246:1     -bind-to none -map-by slot     -x NCCL_DEBUG=INFO -x LD_LIBRARY_PATH     python3 keras_mnist_advanced.py 
execve("/usr/local/bin/mpirun", ["mpirun", "-np", "2", "-H", "192.168.2.243:1,192.168.3.246:1", "-bind-to", "none", "-map-by", "slot", "-x", "NCCL_DEBUG=INFO", "-x", "LD_LIBRARY_PATH", "python3", "keras_mnist_advanced.py"], [/* 33 vars */]) = 0
brk(NULL)                               = 0x176c000
access("/etc/ld.so.nohwcap", F_OK)      = -1 ENOENT (No such file or directory)
mmap(NULL, 8192, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_ANONYMOUS, -1, 0) = 0x7fda47570000
access("/etc/ld.so.preload", R_OK)      = -1 ENOENT (No such file or directory)
open("/usr/local/cuda-8.0/lib64/tls/x86_64/libopen-rte.so.40", O_RDONLY|O_CLOEXEC) = -1 ENOENT (No such file or directory)
stat("/usr/local/cuda-8.0/lib64/tls/x86_64", 0x7fff4ff675c0) = -1 ENOENT (No such file or directory)
open("/usr/local/cuda-8.0/lib64/tls/libopen-rte.so.40", O_RDONLY|O_CLO

BIGBALLON

"""
modfied from MorvanZhou' code!

Know more, visit my Python tutorial page: https://morvanzhou.github.io/tutorials/
My Youtube Channel: https://www.youtube.com/user/MorvanZhou
More about Reinforcement learning: https://morvanzhou.github.io/tutorials/machine-learning/reinforcement-learning/

Dependencies:
tensorflow: 1.1.0
matplotlib