小编Ale*_*sen的帖子

我在Ubuntu 14.04上安装了GPU版本的tensorflow.

我在GPU服务器上,tensorflow可以访问可用的GPU.

我想在CPU上运行tensorflow.

通常情况下,我可以使用env CUDA_VISIBLE_DEVICES=0GPU上运行.0.

如何在CPU之间进行选择？

我没有兴趣重写我的代码 with tf.device("/cpu:0"):

如何解释TensorFlow输出以在GPGPU上构建和执行计算图？

给定以下命令,使用python API执行任意tensorflow脚本.

python3 tensorflow_test.py> out

第一部分stream_executor似乎是它的加载依赖.

I tensorflow/stream_executor/dso_loader.cc:105] successfully opened CUDA library libcublas.so locally
I tensorflow/stream_executor/dso_loader.cc:105] successfully opened CUDA library libcudnn.so locally
I tensorflow/stream_executor/dso_loader.cc:105] successfully opened CUDA library libcufft.so locally
I tensorflow/stream_executor/dso_loader.cc:105] successfully opened CUDA library libcuda.so.1 locally
I tensorflow/stream_executor/dso_loader.cc:105] successfully opened CUDA library libcurand.so locally

什么是NUMA节点？

I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:900] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero

我认为这是它找到可用GPU的时候

I …

Ubuntu 14.04，CUDA版本7.5.18，每晚构建的Tensorflow

tf.nn.max_pool()在tensorflow中运行操作时，出现以下错误：

[tensorflow / stream_executor / cuda / cuda_dnn.cc：286]已加载的cudnn库：5005，但是源代码是针对4007编译的。如果使用二进制安装，请升级您的cudnn库以使其匹配。如果从源构建，请确保加载的库与您在编译配置期间指定的版本匹配。

W tensorflow / stream_executor / stream.cc：577]尝试使用不具有DNN支持的StreamExecutor执行DNN操作

追溯（最近一次通话）：

...

如何在tensorflow的编译配置中指定我的cudnn版本？

如何在 TensorFlow 中重用变量？我想重用tf.contrib.layers.linear

with tf.variable_scope("root") as varscope:
    inputs_1 = tf.constant(0.5, shape=[2, 3, 4])
    inputs_2 = tf.constant(0.5, shape=[2, 3, 4])
    outputs_1 = tf.contrib.layers.linear(inputs_1, 5)
    varscope.reuse_variables()
    outputs_2 = tf.contrib.layers.linear(inputs_2, 5)

但它给了我以下结果

---------------------------------------------------------------------------
ValueError                                Traceback (most recent call last)
<ipython-input-51-a40b9ec68e25> in <module>()
      5     outputs_1 = tf.contrib.layers.linear(inputs_1, 5)
      6     varscope.reuse_variables()
----> 7     outputs_2 = tf.contrib.layers.linear(inputs_2, 5)
...
ValueError: Variable root/fully_connected_1/weights does not exist, or was not created with tf.get_variable(). Did you mean to set reuse=None in VarScope?

以下:scrapy的教程我做了一个简单的图像爬虫(刮掉Bugattis的图像).这在下面的实施例中说明.

但是,按照指南给我留下了一个不起作用的爬虫!它找到所有网址,但不下载图片.

我找到了一个鸭子胶带解决方案:替换ITEM_PIPELINES等等IMAGES_STORE;

ITEM_PIPELINES['scrapy.pipeline.images.FilesPipeline'] = 1 和

IMAGES_STORE - > FILES_STORE

但我不知道为什么会这样呢？我想使用scrapy记录的ImagePipeline.

例

settings.py

BOT_NAME = 'imagespider'
SPIDER_MODULES = ['imagespider.spiders']
NEWSPIDER_MODULE = 'imagespider.spiders'
ITEM_PIPELINES = {
    'scrapy.pipelines.images.ImagesPipeline': 1,
}
IMAGES_STORE = "/home/user/Desktop/imagespider/output"

items.py

import scrapy

class ImageItem(scrapy.Item):
    file_urls = scrapy.Field()
    files = scrapy.Field()

imagespider.py

from imagespider.items import ImageItem
import scrapy


class ImageSpider(scrapy.Spider):
    name = "imagespider"

    start_urls = (
        "https://www.find.com/search=bugatti+veyron",
    )

    def parse(self, response):
        for elem in response.xpath("//img"):
            img_url = elem.xpath("@src").extract_first()
            yield ImageItem(file_urls=[img_url])

我是C++/CUDA 的新手。我尝试通过递归内核的输出（在内核包装器中）来实现并行算法“ reduce ”，它能够处理任何类型的输入大小和线程大小，而不会增加渐近并行运行时间。

>> nvcc -o mycode mycode.cu
>> ./mycode
Segmentail fault.

#include <stdio.h>

/* -------- KERNEL -------- */
__global__ void reduce_kernel(float * d_out, float * d_in, const int size)
{
  // position and threadId
  int pos = blockIdx.x * blockDim.x + threadIdx.x;
  int tid = threadIdx.x;

  // do reduction in global memory
  for (unsigned …