ruoneo
/
blog
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity
You can not select more than 25 topics Topics must start with a letter or number, can include dashes ('-') and can be up to 35 characters long.
4 Commits
1 Branch
0 Tags
18 MiB
 
 
 
Tag: Branch: Tree: main
Branches Tags
${ item.name }
Create tag ${ searchTerm }
Create branch ${ searchTerm }
from 'main'
${ noResults }
blog/public/2021/07/23/Tensorflow中的GPU配置及内存分配/index.html

308 lines
32 KiB
Raw Permalink Blame History

<!DOCTYPE html><html lang="zh-CN" data-theme="light"><head><meta charset="UTF-8"><meta http-equiv="X-UA-Compatible" content="IE=edge"><meta name="viewport" content="width=device-width, initial-scale=1.0, maximum-scale=1.0, user-scalable=no"><title>Tensorflow中的GPU配置及内存分配 | 水榭听泉</title><meta name="author" content="Ruoneo"><meta name="copyright" content="Ruoneo"><meta name="format-detection" content="telephone=no"><meta name="theme-color" content="#ffffff"><meta name="description" content="Nvidia-smi nvidia-sim简称NVSMI，提供监控GPU使用情况和更改GPU状态的功能。这个工具是N卡驱动附带的。smi(System management interface)。 更详细的介绍：https:&#x2F;&#x2F;blog.csdn.net&#x2F;C_chuxin&#x2F;article&#x2F;details&#x2F;82993350 CMD中输入: 1nvidia-smi GPU：本机中的GPU编号（有">
<meta property="og:type" content="article">
<meta property="og:title" content="Tensorflow中的GPU配置及内存分配">
<meta property="og:url" content="https://ismaxu.com/2021/07/23/Tensorflow%E4%B8%AD%E7%9A%84GPU%E9%85%8D%E7%BD%AE%E5%8F%8A%E5%86%85%E5%AD%98%E5%88%86%E9%85%8D/index.html">
<meta property="og:site_name" content="水榭听泉">
<meta property="og:description" content="Nvidia-smi nvidia-sim简称NVSMI，提供监控GPU使用情况和更改GPU状态的功能。这个工具是N卡驱动附带的。smi(System management interface)。 更详细的介绍：https:&#x2F;&#x2F;blog.csdn.net&#x2F;C_chuxin&#x2F;article&#x2F;details&#x2F;82993350 CMD中输入: 1nvidia-smi GPU：本机中的GPU编号（有">
<meta property="og:locale" content="zh_CN">
<meta property="og:image" content="https://i.loli.net/2020/05/01/gkihqEjXxJ5UZ1C.jpg">
<meta property="article:published_time" content="2021-07-23T01:01:59.000Z">
<meta property="article:modified_time" content="2021-11-06T05:54:49.000Z">
<meta property="article:author" content="Ruoneo">
<meta name="twitter:card" content="summary">
<meta name="twitter:image" content="https://i.loli.net/2020/05/01/gkihqEjXxJ5UZ1C.jpg"><link rel="shortcut icon" href="/img/avatar.jpg"><link rel="canonical" href="https://ismaxu.com/2021/07/23/Tensorflow%E4%B8%AD%E7%9A%84GPU%E9%85%8D%E7%BD%AE%E5%8F%8A%E5%86%85%E5%AD%98%E5%88%86%E9%85%8D/"><link rel="preconnect" href="//cdn.jsdelivr.net"/><link rel="stylesheet" href="/css/index.css"><link rel="stylesheet" href="https://cdn.jsdelivr.net/npm/@fortawesome/fontawesome-free/css/all.min.css" media="print" onload="this.media='all'"><link rel="stylesheet" href="https://cdn.jsdelivr.net/npm/@fancyapps/ui/dist/fancybox.min.css" media="print" onload="this.media='all'"><script>const GLOBAL_CONFIG = {
root: '/',
algolia: undefined,
localSearch: {"path":"/search.xml","preload":false,"languages":{"hits_empty":"找不到您查询的内容：${query}"}},
translate: undefined,
noticeOutdate: undefined,
highlight: {"plugin":"highlighjs","highlightCopy":true,"highlightLang":true,"highlightHeightLimit":false},
copy: {
success: '复制成功',
error: '复制错误',
noSupport: '浏览器不支持'
},
relativeDate: {
homepage: false,
post: false
},
runtime: '',
date_suffix: {
just: '刚刚',
min: '分钟前',
hour: '小时前',
day: '天前',
month: '个月前'
},
copyright: undefined,
lightbox: 'fancybox',
Snackbar: undefined,
source: {
justifiedGallery: {
js: 'https://cdn.jsdelivr.net/npm/flickr-justified-gallery/dist/fjGallery.min.js',
css: 'https://cdn.jsdelivr.net/npm/flickr-justified-gallery/dist/fjGallery.min.css'
}
},
isPhotoFigcaption: false,
islazyload: true,
isAnchor: false
}</script><script id="config-diff">var GLOBAL_CONFIG_SITE = {
title: 'Tensorflow中的GPU配置及内存分配',
isPost: true,
isHome: false,
isHighlightShrink: false,
isToc: true,
postUpdate: '2021-11-06 13:54:49'
}</script><noscript><style type="text/css">
#nav {
opacity: 1
}
.justified-gallery img {
opacity: 1
}
#recent-posts time,
#post-meta time {
display: inline !important
}
</style></noscript><script>(win=>{
win.saveToLocal = {
set: function setWithExpiry(key, value, ttl) {
if (ttl === 0) return
const now = new Date()
const expiryDay = ttl * 86400000
const item = {
value: value,
expiry: now.getTime() + expiryDay,
}
localStorage.setItem(key, JSON.stringify(item))
},
get: function getWithExpiry(key) {
const itemStr = localStorage.getItem(key)
if (!itemStr) {
return undefined
}
const item = JSON.parse(itemStr)
const now = new Date()
if (now.getTime() > item.expiry) {
localStorage.removeItem(key)
return undefined
}
return item.value
}
}
win.getScript = url => new Promise((resolve, reject) => {
const script = document.createElement('script')
script.src = url
script.async = true
script.onerror = reject
script.onload = script.onreadystatechange = function() {
const loadState = this.readyState
if (loadState && loadState !== 'loaded' && loadState !== 'complete') return
script.onload = script.onreadystatechange = null
resolve()
}
document.head.appendChild(script)
})
win.activateDarkMode = function () {
document.documentElement.setAttribute('data-theme', 'dark')
if (document.querySelector('meta[name="theme-color"]') !== null) {
document.querySelector('meta[name="theme-color"]').setAttribute('content', '#0d0d0d')
}
}
win.activateLightMode = function () {
document.documentElement.setAttribute('data-theme', 'light')
if (document.querySelector('meta[name="theme-color"]') !== null) {
document.querySelector('meta[name="theme-color"]').setAttribute('content', '#ffffff')
}
}
const t = saveToLocal.get('theme')
if (t === 'dark') activateDarkMode()
else if (t === 'light') activateLightMode()
const asideStatus = saveToLocal.get('aside-status')
if (asideStatus !== undefined) {
if (asideStatus === 'hide') {
document.documentElement.classList.add('hide-aside')
} else {
document.documentElement.classList.remove('hide-aside')
}
}
const detectApple = () => {
if(/iPad|iPhone|iPod|Macintosh/.test(navigator.userAgent)){
document.documentElement.classList.add('apple')
}
}
detectApple()
})(window)</script><meta name="generator" content="Hexo 6.2.0"></head><body><div id="web_bg"></div><div id="sidebar"><div id="menu-mask"></div><div id="sidebar-menus"><div class="avatar-img is-center"><img src="/img/avatar.jpg" onerror="onerror=null;src='/img/friend_404.gif'" alt="avatar"/></div><div class="sidebar-site-data site-data is-center"><a href="/archives/"><div class="headline">文章</div><div class="length-num">64</div></a><a href="/tags/"><div class="headline">标签</div><div class="length-num">26</div></a><a href="/categories/"><div class="headline">分类</div><div class="length-num">4</div></a></div><hr/><div class="menus_items"><div class="menus_item"><a class="site-page" href="/"><i class="fa-fw fas fa-home"></i><span> 首页</span></a></div><div class="menus_item"><a class="site-page" href="/archives/"><i class="fa-fw fas fa-archive"></i><span> 时间轴</span></a></div><div class="menus_item"><a class="site-page" href="/tags/"><i class="fa-fw fas fa-tags"></i><span> 标签</span></a></div><div class="menus_item"><a class="site-page" href="/categories/"><i class="fa-fw fas fa-folder-open"></i><span> 分类</span></a></div><div class="menus_item"><a class="site-page" href="/messageboard/"><i class="fa-fw fas fa-comment-dots"></i><span> 留言板</span></a></div><div class="menus_item"><a class="site-page" href="/links/"><i class="fa-fw fas fa-link"></i><span> 友链</span></a></div><div class="menus_item"><a class="site-page" href="/about/"><i class="fa-fw fas fa-heart"></i><span> 关于</span></a></div></div></div></div><div class="post" id="body-wrap"><header class="post-bg" id="page-header" style="background-image: url('https://i.loli.net/2020/05/01/gkihqEjXxJ5UZ1C.jpg')"><nav id="nav"><span id="blog_name"><a id="site-name" href="/">水榭听泉</a></span><div id="menus"><div id="search-button"><a class="site-page social-icon search"><i class="fas fa-search fa-fw"></i><span> 搜索</span></a></div><div class="menus_items"><div class="menus_item"><a class="site-page" href="/"><i class="fa-fw fas fa-home"></i><span> 首页</span></a></div><div class="menus_item"><a class="site-page" href="/archives/"><i class="fa-fw fas fa-archive"></i><span> 时间轴</span></a></div><div class="menus_item"><a class="site-page" href="/tags/"><i class="fa-fw fas fa-tags"></i><span> 标签</span></a></div><div class="menus_item"><a class="site-page" href="/categories/"><i class="fa-fw fas fa-folder-open"></i><span> 分类</span></a></div><div class="menus_item"><a class="site-page" href="/messageboard/"><i class="fa-fw fas fa-comment-dots"></i><span> 留言板</span></a></div><div class="menus_item"><a class="site-page" href="/links/"><i class="fa-fw fas fa-link"></i><span> 友链</span></a></div><div class="menus_item"><a class="site-page" href="/about/"><i class="fa-fw fas fa-heart"></i><span> 关于</span></a></div></div><div id="toggle-menu"><a class="site-page"><i class="fas fa-bars fa-fw"></i></a></div></div></nav><div id="post-info"><h1 class="post-title">Tensorflow中的GPU配置及内存分配<a class="post-edit-link" href="https://git.ismaxu.com/ruoneo/hexo-generator/src/branch/main/source/_posts/Tensorflow中的GPU配置及内存分配.md" title="编辑" target="_blank"><i class="fas fa-pencil-alt"></i></a></h1><div id="post-meta"><div class="meta-firstline"><span class="post-meta-date"><i class="far fa-calendar-alt fa-fw post-meta-icon"></i><span class="post-meta-label">发表于</span><time class="post-meta-date-created" datetime="2021-07-23T01:01:59.000Z" title="发表于 2021-07-23 09:01:59">2021-07-23</time><span class="post-meta-separator">|</span><i class="fas fa-history fa-fw post-meta-icon"></i><span class="post-meta-label">更新于</span><time class="post-meta-date-updated" datetime="2021-11-06T05:54:49.000Z" title="更新于 2021-11-06 13:54:49">2021-11-06</time></span></div><div class="meta-secondline"><span class="post-meta-separator">|</span><span class="post-meta-wordcount"><i class="far fa-file-word fa-fw post-meta-icon"></i><span class="post-meta-label">字数总计:</span><span class="word-count">787</span><span class="post-meta-separator">|</span><i class="far fa-clock fa-fw post-meta-icon"></i><span class="post-meta-label">阅读时长:</span><span>3分钟</span></span></div></div></div></header><main class="layout" id="content-inner"><div id="post"><article class="post-content" id="article-container"><h2 id="Nvidia-smi">Nvidia-smi</h2>
<p>nvidia-sim简称NVSMI，提供监控GPU使用情况和更改GPU状态的功能。这个工具是N卡驱动附带的。smi(System management interface)。</p>
<p>更详细的介绍：<a target="_blank" rel="noopener" href="https://blog.csdn.net/C_chuxin/article/details/82993350">https://blog.csdn.net/C_chuxin/article/details/82993350</a></p>
<p>CMD中输入:</p>
<figure class="highlight text"><table><tr><td class="gutter"><pre><span class="line">1</span><br></pre></td><td class="code"><pre><span class="line">nvidia-smi</span><br></pre></td></tr></table></figure>
<p><img src="../images/5bc51343e21394d4f82ce4a63916e4af7df2acf0be93a444928e300bd564a089.png" alt="图 2"></p>
<ul>
<li>
<p>GPU：本机中的GPU编号（有多块显卡的时候，从0开始编号）图上GPU的编号是：0</p>
</li>
<li>
<p>Fan：风扇转速（0%-100%），N/A表示没有风扇</p>
</li>
<li>
<p>Name：GPU类型，图上GPU的类型是：Tesla T4</p>
</li>
<li>
<p>Temp：GPU的温度（GPU温度过高会导致GPU的频率下降）</p>
</li>
<li>
<p>Perf：GPU的性能状态，从P0（最大性能）到P12（最小性能），图上是：P0</p>
</li>
<li>
<p>Persistence-M：持续模式的状态，持续模式虽然耗能大，但是在新的GPU应用启动时花费的时间更少，图上显示的是：off</p>
</li>
<li>
<p>Pwr：Usager/Cap：能耗表示，Usage：用了多少，Cap总共多少</p>
</li>
<li>
<p>Bus-Id：GPU总线相关显示，domain：bus：device.function</p>
</li>
<li>
<p>Disp.A：Display Active ，表示GPU的显示是否初始化</p>
</li>
<li>
<p>Memory-Usage：显存使用率</p>
</li>
<li>
<p>Volatile GPU-Util：GPU使用率</p>
</li>
<li>
<p>Uncorr. ECC：关于ECC的东西，是否开启错误检查和纠正技术，0/disabled,1/enabled</p>
</li>
<li>
<p>Compute M：计算模式，0/DEFAULT,1/EXCLUSIVE_PROCESS,2/PROHIBITED</p>
</li>
<li>
<p>Processes：显示每个进程占用的显存使用率、进程号、占用的哪个GPU</p>
</li>
</ul>
<p>隔几秒刷新一下显存状态：nvidia-smi -l 秒数</p>
<p>隔两秒刷新一下GPU的状态：nvidia-smi -l 2</p>
<h2 id="tensorflow的显卡使用方式">tensorflow的显卡使用方式</h2>
<h3 id="1、直接使用">1、直接使用</h3>
<p>这种方式会把当前机器上所有的显卡的剩余显存基本都占用，注意是机器上所有显卡的剩余显存。</p>
<figure class="highlight python"><table><tr><td class="gutter"><pre><span class="line">1</span><br><span class="line">2</span><br><span class="line">3</span><br><span class="line">4</span><br><span class="line">5</span><br><span class="line">6</span><br><span class="line">7</span><br><span class="line">8</span><br><span class="line">9</span><br><span class="line">10</span><br><span class="line">11</span><br><span class="line">12</span><br></pre></td><td class="code"><pre><span class="line"><span class="keyword">with</span> tf.compat.v1.Session() <span class="keyword">as</span> sess:</span><br><span class="line"> <span class="comment"># 输入图片为256x256，2个分类</span></span><br><span class="line"> shape, classes = (<span class="number">224</span>, <span class="number">224</span>, <span class="number">3</span>), <span class="number">20</span></span><br><span class="line"> <span class="comment"># 调用keras的ResNet50模型</span></span><br><span class="line"> model = keras.applications.resnet50.ResNet50(input_shape = shape, weights=<span class="literal">None</span>, classes=classes)</span><br><span class="line"> model.<span class="built_in">compile</span>(optimizer=<span class="string">&quot;adam&quot;</span>, loss=<span class="string">&quot;sparse_categorical_crossentropy&quot;</span>, metrics=[<span class="string">&quot;accuracy&quot;</span>])</span><br><span class="line"></span><br><span class="line"> <span class="comment"># 训练模型 categorical_crossentropy sparse_categorical_crossentropy</span></span><br><span class="line"> <span class="comment"># training = model.fit(train_x, train_y, epochs=50, batch_size=10)</span></span><br><span class="line"> model.fit(train_x,train_y,validation_data=(test_x, test_y), epochs=<span class="number">20</span>, batch_size=<span class="number">6</span>,verbose=<span class="number">2</span>)</span><br><span class="line"> <span class="comment"># # 把训练好的模型保存到文件</span></span><br><span class="line"> model.save(<span class="string">&#x27;resnet_model_dog_n_face.h5&#x27;</span>)</span><br></pre></td></tr></table></figure>
<h3 id="2、分配比例使用">2、分配比例使用</h3>
<figure class="highlight python"><table><tr><td class="gutter"><pre><span class="line">1</span><br><span class="line">2</span><br><span class="line">3</span><br><span class="line">4</span><br><span class="line">5</span><br></pre></td><td class="code"><pre><span class="line"><span class="keyword">from</span> tensorflow.compat.v1 <span class="keyword">import</span> ConfigProto<span class="comment"># tf 2.x的写法</span></span><br><span class="line">config =ConfigProto()</span><br><span class="line">config.gpu_options.per_process_gpu_memory_fraction=<span class="number">0.6</span></span><br><span class="line"><span class="keyword">with</span> tf.compat.v1.Session(config=config) <span class="keyword">as</span> sess:</span><br><span class="line"> model = keras.applications.resnet50.ResNet50(input_shape = shape, weights=<span class="literal">None</span>, classes=classes)</span><br></pre></td></tr></table></figure>
<h3 id="3-动态申请使用">3. 动态申请使用</h3>
<p><strong>这种方式是动态申请显存的，只会申请内存，不会释放内存。而且如果别人的程序把剩余显卡全部占了，就会报错。</strong></p>
<figure class="highlight python"><table><tr><td class="gutter"><pre><span class="line">1</span><br><span class="line">2</span><br><span class="line">3</span><br><span class="line">4</span><br><span class="line">5</span><br></pre></td><td class="code"><pre><span class="line">config = tf.compat.v1.ConfigProto()</span><br><span class="line">config.gpu_options.allow_growth = <span class="literal">True</span></span><br><span class="line">session = tf.compat.v1.InteractiveSession(config=config)</span><br><span class="line"><span class="keyword">with</span> tf.compat.v1.Session(config=config) <span class="keyword">as</span> sess:</span><br><span class="line"> model</span><br></pre></td></tr></table></figure>
<h3 id="4-指定GPU">4 指定GPU</h3>
<ul>
<li>在有多块GPU的服务器上运行tensorflow的时候，如果使用python编程，则可指定GPU，代码如下：</li>
</ul>
<figure class="highlight python"><table><tr><td class="gutter"><pre><span class="line">1</span><br><span class="line">2</span><br><span class="line">3</span><br><span class="line">4</span><br><span class="line">5</span><br><span class="line">6</span><br><span class="line">7</span><br><span class="line">8</span><br><span class="line">9</span><br><span class="line">10</span><br></pre></td><td class="code"><pre><span class="line"><span class="keyword">import</span> os</span><br><span class="line">os.environ[<span class="string">&quot;CUDA_VISIBLE_DEVICES&quot;</span>] = <span class="string">&quot;2&quot;</span></span><br><span class="line"></span><br><span class="line">os.environ[<span class="string">&quot;CUDA_VISIBLE_DEVICES&quot;</span>]=<span class="string">&quot;-1&quot;</span> CPU模式</span><br><span class="line"></span><br><span class="line"></span><br><span class="line">CUDA_VISIBLE_DEVICES=<span class="number">1</span> Only device <span class="number">1</span> will be seen</span><br><span class="line">CUDA_VISIBLE_DEVICES=<span class="number">0</span>,<span class="number">1</span> Devices <span class="number">0</span> <span class="keyword">and</span> <span class="number">1</span> will be visible</span><br><span class="line">CUDA_VISIBLE_DEVICES=<span class="string">&quot;0,1&quot;</span> Same <span class="keyword">as</span> above, quotation marks are optional</span><br><span class="line">CUDA_VISIBLE_DEVICES=<span class="number">0</span>,<span class="number">2</span>,<span class="number">3</span> Devices <span class="number">0</span>, <span class="number">2</span>, <span class="number">3</span> will be visible; device <span class="number">1</span> <span class="keyword">is</span> masked</span><br></pre></td></tr></table></figure>
<ul>
<li>另一种方案-给指定的Session分配单独的GPU</li>
</ul>
<figure class="highlight text"><table><tr><td class="gutter"><pre><span class="line">1</span><br><span class="line">2</span><br><span class="line">3</span><br></pre></td><td class="code"><pre><span class="line">用 tf.device(&#x27;/gpu:1&#x27;) 指定Session在第二块GPU上运行。 </span><br><span class="line">tensorflow中不同的GPU使用`/gpu:0`和`/gpu:1`区分，</span><br><span class="line">CPU不区分设备号，表示为`/cpu:0`。</span><br></pre></td></tr></table></figure>
<blockquote>
<p>未完待续…</p>
</blockquote>
</article><div class="post-copyright"><div class="post-copyright__author"><span class="post-copyright-meta">文章作者: </span><span class="post-copyright-info"><a href="https://ismaxu.com">Ruoneo</a></span></div><div class="post-copyright__type"><span class="post-copyright-meta">文章链接: </span><span class="post-copyright-info"><a href="https://ismaxu.com/2021/07/23/Tensorflow%E4%B8%AD%E7%9A%84GPU%E9%85%8D%E7%BD%AE%E5%8F%8A%E5%86%85%E5%AD%98%E5%88%86%E9%85%8D/">https://ismaxu.com/2021/07/23/Tensorflow%E4%B8%AD%E7%9A%84GPU%E9%85%8D%E7%BD%AE%E5%8F%8A%E5%86%85%E5%AD%98%E5%88%86%E9%85%8D/</a></span></div><div class="post-copyright__notice"><span class="post-copyright-meta">版权声明: </span><span class="post-copyright-info">本博客所有文章除特别声明外，均采用 <a href="https://creativecommons.org/licenses/by-nc-sa/4.0/" target="_blank">CC BY-NC-SA 4.0</a> 许可协议。转载请注明来自 <a href="https://ismaxu.com" target="_blank">水榭听泉</a>！</span></div></div><div class="tag_share"><div class="post-meta__tag-list"></div><div class="post_share"><div class="social-share" data-image="https://i.loli.net/2020/05/01/gkihqEjXxJ5UZ1C.jpg" data-sites="facebook,twitter,wechat,weibo,qq"></div><link rel="stylesheet" href="https://cdn.jsdelivr.net/npm/butterfly-extsrc/sharejs/dist/css/share.min.css" media="print" onload="this.media='all'"><script src="https://cdn.jsdelivr.net/npm/butterfly-extsrc/sharejs/dist/js/social-share.min.js" defer></script></div></div><nav class="pagination-post" id="pagination"><div class="prev-post pull-left"><a href="/2021/07/31/Graph%20Attentional%20Layer/"><img class="prev-cover" src="https://i.loli.net/2020/05/01/gkihqEjXxJ5UZ1C.jpg" onerror="onerror=null;src='/img/404.jpg'" alt="cover of previous post"><div class="pagination-info"><div class="label">上一篇</div><div class="prev_info">Graph Attention Networks (图注意力网络)</div></div></a></div><div class="next-post pull-right"><a href="/2021/07/13/Drug%E2%80%93drug%20interaction%20prediction%20with%20Wasserstein%20Adversarial%20Autoencoder-based%20knowledge%20graph%20embeddings/"><img class="next-cover" src="https://i.loli.net/2020/05/01/gkihqEjXxJ5UZ1C.jpg" onerror="onerror=null;src='/img/404.jpg'" alt="cover of next post"><div class="pagination-info"><div class="label">下一篇</div><div class="next_info">Drug–drug interaction prediction with Wasserstein Adversarial Autoencoder-based knowledge graph embeddings</div></div></a></div></nav><hr/><div id="post-comment"><div class="comment-head"><div class="comment-headline"><i class="fas fa-comments fa-fw"></i><span> 评论</span></div><div id="comment-switch"><span class="first-comment">Twikoo</span><span class="switch-btn"></span><span class="second-comment">Gitalk</span></div></div><div class="comment-wrap"><div><div id="twikoo-wrap"></div></div><div><div id="gitalk-container"></div></div></div></div></div><div class="aside-content" id="aside-content"><div class="card-widget card-info"><div class="is-center"><div class="avatar-img"><img src="/img/avatar.jpg" onerror="this.onerror=null;this.src='/img/friend_404.gif'" alt="avatar"/></div><div class="author-info__name">Ruoneo</div><div class="author-info__description">嗨! 你好!</div></div><div class="card-info-data site-data is-center"><a href="/archives/"><div class="headline">文章</div><div class="length-num">64</div></a><a href="/tags/"><div class="headline">标签</div><div class="length-num">26</div></a><a href="/categories/"><div class="headline">分类</div><div class="length-num">4</div></a></div><div class="card-info-social-icons is-center"><a class="social-icon" href="https://github.com/ruoneo" target="_blank" title="Github"><i class="fab fa-github"></i></a><a class="social-icon" href="mailto:xum321@gmail.com" target="_blank" title="Email"><i class="fas fa-envelope"></i></a></div></div><div class="card-widget card-announcement"><div class="item-headline"><i class="fas fa-bullhorn fa-shake"></i><span>公告</span></div><div class="announcement_content">感谢来访。希望这里有对你有帮助的内容。</div></div><div class="sticky_layout"><div class="card-widget" id="card-toc"><div class="item-headline"><i class="fas fa-stream"></i><span>目录</span><span class="toc-percentage"></span></div><div class="toc-content"><ol class="toc"><li class="toc-item toc-level-2"><a class="toc-link" href="#Nvidia-smi"><span class="toc-text">Nvidia-smi</span></a></li><li class="toc-item toc-level-2"><a class="toc-link" href="#tensorflow%E7%9A%84%E6%98%BE%E5%8D%A1%E4%BD%BF%E7%94%A8%E6%96%B9%E5%BC%8F"><span class="toc-text">tensorflow的显卡使用方式</span></a><ol class="toc-child"><li class="toc-item toc-level-3"><a class="toc-link" href="#1%E3%80%81%E7%9B%B4%E6%8E%A5%E4%BD%BF%E7%94%A8"><span class="toc-text">1、直接使用</span></a></li><li class="toc-item toc-level-3"><a class="toc-link" href="#2%E3%80%81%E5%88%86%E9%85%8D%E6%AF%94%E4%BE%8B%E4%BD%BF%E7%94%A8"><span class="toc-text">2、分配比例使用</span></a></li><li class="toc-item toc-level-3"><a class="toc-link" href="#3-%E5%8A%A8%E6%80%81%E7%94%B3%E8%AF%B7%E4%BD%BF%E7%94%A8"><span class="toc-text">3. 动态申请使用</span></a></li><li class="toc-item toc-level-3"><a class="toc-link" href="#4-%E6%8C%87%E5%AE%9AGPU"><span class="toc-text">4 指定GPU</span></a></li></ol></li></ol></div></div><div class="card-widget card-recent-post"><div class="item-headline"><i class="fas fa-history"></i><span>最新文章</span></div><div class="aside-list"><div class="aside-list-item no-cover"><div class="content"><a class="title" href="/2024/07/06/git%E5%B1%80%E5%9F%9F%E7%BD%91%E9%83%A8%E7%BD%B2%E6%96%B9%E6%B3%95/" title="git局域网部署方法">git局域网部署方法</a><time datetime="2024-07-06T08:13:10.000Z" title="发表于 2024-07-06 16:13:10">2024-07-06</time></div></div><div class="aside-list-item no-cover"><div class="content"><a class="title" href="/2023/09/25/%E5%85%B3%E4%BA%8EPyQt%E7%9A%84connect%E4%BC%A0%E5%85%A5lambda%E5%87%BD%E6%95%B0%E4%BD%9C%E4%B8%BA%E5%8F%82%E6%95%B0%E8%B0%83%E7%94%A8%E6%A7%BD%E5%87%BD%E6%95%B0%E7%9A%84%E4%B8%80%E7%82%B9%E6%80%9D%E8%80%83/" title="关于PyQt的connect传入lambda函数作为参数调用槽函数的一点思考">关于PyQt的connect传入lambda函数作为参数调用槽函数的一点思考</a><time datetime="2023-09-25T13:53:35.000Z" title="发表于 2023-09-25 21:53:35">2023-09-25</time></div></div><div class="aside-list-item no-cover"><div class="content"><a class="title" href="/2022/03/30/%E5%A4%8D%E4%B9%A05/" title="Spring注解">Spring注解</a><time datetime="2022-03-30T07:22:16.000Z" title="发表于 2022-03-30 15:22:16">2022-03-30</time></div></div><div class="aside-list-item no-cover"><div class="content"><a class="title" href="/2022/03/30/%E5%A4%8D%E4%B9%A04/" title="Spring">Spring</a><time datetime="2022-03-30T02:23:10.000Z" title="发表于 2022-03-30 10:23:10">2022-03-30</time></div></div><div class="aside-list-item no-cover"><div class="content"><a class="title" href="/2022/03/22/MultiDTI_%20drug%E2%80%93target%20interaction%20prediction%20based%20on%20multi-modal%20representation%20learning%20to%20bridge%20the%20gap%20between%20new%20chemical/" title="MultiDTI_ drug–target interaction prediction based on multi-modal representation learning to bridge the gap between new chemical">MultiDTI_ drug–target interaction prediction based on multi-modal representation learning to bridge the gap between new chemical</a><time datetime="2022-03-22T06:29:57.000Z" title="发表于 2022-03-22 14:29:57">2022-03-22</time></div></div></div></div></div></div></main><footer id="footer" style="background: color(orange)"><div id="footer-wrap"><div class="copyright">&copy;2024 By Ruoneo</div><div class="footer_custom_text"><div><a target="_blank" rel="noopener" href="https://beian.miit.gov.cn"><img class="icp-icon"><span>豫ICP备 2022020326号</span></a>&nbsp|&nbsp<a target="_blank" rel="noopener" href="http://www.beian.gov.cn/portal/registerSystemInfo?recordcode=41020202000143"><span>豫公网安备 41020202000143号</span></a></div></div></div></footer></div><div id="rightside"><div id="rightside-config-hide"><button id="readmode" type="button" title="阅读模式"><i class="fas fa-book-open"></i></button><button id="darkmode" type="button" title="浅色和深色模式转换"><i class="fas fa-adjust"></i></button><button id="hide-aside-btn" type="button" title="单栏和双栏切换"><i class="fas fa-arrows-alt-h"></i></button></div><div id="rightside-config-show"><button id="rightside_config" type="button" title="设置"><i class="fas fa-cog fa-spin"></i></button><button class="close" id="mobile-toc-button" type="button" title="目录"><i class="fas fa-list-ul"></i></button><a id="to_comment" href="#post-comment" title="直达评论"><i class="fas fa-comments"></i></a><button id="go-up" type="button" title="回到顶部"><i class="fas fa-arrow-up"></i></button></div></div><div id="local-search"><div class="search-dialog"><nav class="search-nav"><span class="search-dialog-title">搜索</span><span id="loading-status"></span><button class="search-close-button"><i class="fas fa-times"></i></button></nav><div class="is-center" id="loading-database"><i class="fas fa-spinner fa-pulse"></i><span> 数据库加载中</span></div><div class="search-wrap"><div id="local-search-input"><div class="local-search-box"><input class="local-search-box--input" placeholder="搜索文章" type="text"/></div></div><hr/><div id="local-search-results"></div></div></div><div id="search-mask"></div></div><div><script src="/js/utils.js"></script><script src="/js/main.js"></script><script src="https://cdn.jsdelivr.net/npm/@fancyapps/ui/dist/fancybox.umd.min.js"></script><script src="https://cdn.jsdelivr.net/npm/vanilla-lazyload/dist/lazyload.iife.min.js"></script><script src="/js/search/local-search.js"></script><div class="js-pjax"><script>(()=>{
const init = () => {
twikoo.init(Object.assign({
el: '#twikoo-wrap',
envId: 'https://tk.ismaxu.com',
region: '',
onCommentLoaded: function () {
btf.loadLightbox(document.querySelectorAll('#twikoo .tk-content img:not(.tk-owo-emotion)'))
}
}, null))
}
const getCount = () => {
const countELement = document.getElementById('twikoo-count')
if(!countELement) return
twikoo.getCommentsCount({
envId: 'https://tk.ismaxu.com',
region: '',
urls: [window.location.pathname],
includeReply: false
}).then(function (res) {
countELement.innerText = res[0].count
}).catch(function (err) {
console.error(err);
});
}
const runFn = () => {
init()
}
const loadTwikoo = () => {
if (typeof twikoo === 'object') {
setTimeout(runFn,0)
return
}
getScript('https://cdn.jsdelivr.net/npm/twikoo/dist/twikoo.all.min.js').then(runFn)
}
if ('Twikoo' === 'Twikoo' || !false) {
if (false) btf.loadComment(document.getElementById('twikoo-wrap'), loadTwikoo)
else loadTwikoo()
} else {
window.loadOtherComment = () => {
loadTwikoo()
}
}
})()</script><script>function addGitalkSource () {
const ele = document.createElement('link')
ele.rel = 'stylesheet'
ele.href= 'https://cdn.jsdelivr.net/npm/gitalk/dist/gitalk.min.css'
document.getElementsByTagName('head')[0].appendChild(ele)
}
function loadGitalk () {
function initGitalk () {
var gitalk = new Gitalk(Object.assign({
clientID: 'f9bab17720653ab07b39',
clientSecret: 'e77fbcac5c4043ddf8d845aee7d6d5f118f607e2',
repo: 'ruoneo-comment-repo',
owner: 'ruoneo',
admin: ['ruoneo'],
id: '0b9b1fed7340a32fe34712f80bd7ee30',
updateCountCallback: commentCount
},null))
gitalk.render('gitalk-container')
}
if (typeof Gitalk === 'function') initGitalk()
else {
addGitalkSource()
getScript('https://cdn.jsdelivr.net/npm/gitalk/dist/gitalk.min.js').then(initGitalk)
}
}
function commentCount(n){
let isCommentCount = document.querySelector('#post-meta .gitalk-comment-count')
if (isCommentCount) {
isCommentCount.innerHTML= n
}
}
if ('Twikoo' === 'Gitalk' || !false) {
if (false) btf.loadComment(document.getElementById('gitalk-container'), loadGitalk)
else loadGitalk()
} else {
function loadOtherComment () {
loadGitalk()
}
}</script></div></div></body></html>