yexinru/doc-exports
1
0
Fork 0
forked from docs/doc-exports
Code Pull Requests Activity
Files
fix_python
doc-exports/docs/modelarts/umn/develop-modelarts-0104.html
Lai, Weijian 6aa966a79a ModelArts UMN 24.3.0 version 
Reviewed-by: Pruthi, Vineet <vineet.pruthi@t-systems.com>
Co-authored-by: Lai, Weijian <laiweijian4@huawei.com>
Co-committed-by: Lai, Weijian <laiweijian4@huawei.com>
2024-11-02 09:04:52 +00:00

310 lines
28 KiB
HTML
Raw Permalink Blame History

<a name="EN-US_TOPIC_0000002079097957"></a><a name="EN-US_TOPIC_0000002079097957"></a>
<h1 class="topictitle1">Viewing Environment Variables of a Training Container</h1>
<div id="body0000001394276053"><div class="section" id="EN-US_TOPIC_0000002079097957__section39451442102418"><h4 class="sectiontitle">What Is an Environment Variable</h4><p id="EN-US_TOPIC_0000002079097957__p1544022782718">This section describes environment variables preset in a training container. The environment variables include:</p>
<ul id="EN-US_TOPIC_0000002079097957__ul216142419117"><li id="EN-US_TOPIC_0000002079097957__li161610241615">Path environment variables</li><li id="EN-US_TOPIC_0000002079097957__li18208123919119">Environment variables of a distributed training job</li><li id="EN-US_TOPIC_0000002079097957__li199126457119">Nvidia Collective multi-GPU Communication Library (NCCL) environment variables</li><li id="EN-US_TOPIC_0000002079097957__li11241135311117">OBS environment variables</li><li id="EN-US_TOPIC_0000002079097957__li1680214311123">Environment variables of the pip source</li><li id="EN-US_TOPIC_0000002079097957__li5610841422">Environment variables of the API Gateway address</li><li id="EN-US_TOPIC_0000002079097957__li57059491620">Environment variables of job metadata </li></ul>
</div>
<div class="section" id="EN-US_TOPIC_0000002079097957__section1094115478346"><h4 class="sectiontitle">Configuring Environment Variables</h4><p id="EN-US_TOPIC_0000002079097957__p272055293411">When you create a training job, you can add environment variables or modify environment variables preset in the training container.</p>
<div class="fignone" id="EN-US_TOPIC_0000002079097957__fig01308309426"><span class="figcap"><b>Figure 1 </b>Setting environment variables</span><br><span><img id="EN-US_TOPIC_0000002079097957__image15845195810245" src="figure/en-us_image_0000002079098133.png" width="469.49" height="46.1909" title="Click to enlarge" class="imgResize"></span></div>
</div>
<div class="section" id="EN-US_TOPIC_0000002079097957__section141942015386"><h4 class="sectiontitle">Environment Variables Preset in a Training Container</h4><p id="EN-US_TOPIC_0000002079097957__p1917630401">The following tables list environment variables preset in a training container.</p>
<p id="EN-US_TOPIC_0000002079097957__p129131350306">The environment variable values are examples.</p>
</div>
<div class="tablenoborder"><table cellpadding="4" cellspacing="0" summary="" id="EN-US_TOPIC_0000002079097957__table6926133121913" frame="border" border="1" rules="all"><caption><b>Table 1 </b>Path environment variables</caption><thead align="left"><tr id="EN-US_TOPIC_0000002079097957__row1899515314197"><th align="left" class="cellrowborder" valign="top" width="21.181818181818183%" id="mcps1.3.4.2.4.1.1"><p id="EN-US_TOPIC_0000002079097957__p2099516317196">Variable</p>
</th>
<th align="left" class="cellrowborder" valign="top" width="34.505050505050505%" id="mcps1.3.4.2.4.1.2"><p id="EN-US_TOPIC_0000002079097957__p10995173119190">Description</p>
</th>
<th align="left" class="cellrowborder" valign="top" width="44.31313131313131%" id="mcps1.3.4.2.4.1.3"><p id="EN-US_TOPIC_0000002079097957__p099511319197">Example</p>
</th>
</tr>
</thead>
<tbody><tr id="EN-US_TOPIC_0000002079097957__row139951731181911"><td class="cellrowborder" valign="top" width="21.181818181818183%" headers="mcps1.3.4.2.4.1.1 "><p id="EN-US_TOPIC_0000002079097957__p199563115197">PATH</p>
</td>
<td class="cellrowborder" valign="top" width="34.505050505050505%" headers="mcps1.3.4.2.4.1.2 "><p id="EN-US_TOPIC_0000002079097957__p15995153115197">Executable file paths</p>
</td>
<td class="cellrowborder" valign="top" width="44.31313131313131%" headers="mcps1.3.4.2.4.1.3 "><p id="EN-US_TOPIC_0000002079097957__p13995113117195">PATH=/usr/local/nvidia/bin:/usr/local/cuda/bin:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin</p>
</td>
</tr>
<tr id="EN-US_TOPIC_0000002079097957__row119955315198"><td class="cellrowborder" valign="top" width="21.181818181818183%" headers="mcps1.3.4.2.4.1.1 "><p id="EN-US_TOPIC_0000002079097957__p19958311191">LD_LIBRARY_PATH</p>
</td>
<td class="cellrowborder" valign="top" width="34.505050505050505%" headers="mcps1.3.4.2.4.1.2 "><p id="EN-US_TOPIC_0000002079097957__p899543181913">Dynamic load library paths </p>
</td>
<td class="cellrowborder" valign="top" width="44.31313131313131%" headers="mcps1.3.4.2.4.1.3 "><p id="EN-US_TOPIC_0000002079097957__p169951431141911">LD_LIBRARY_PATH=/usr/local/seccomponent/lib:/usr/local/cuda/lib64:/usr/local/cuda/compat:/root/miniconda3/lib:/usr/local/nvidia/lib:/usr/local/nvidia/lib64</p>
</td>
</tr>
<tr id="EN-US_TOPIC_0000002079097957__row6995183117198"><td class="cellrowborder" valign="top" width="21.181818181818183%" headers="mcps1.3.4.2.4.1.1 "><p id="EN-US_TOPIC_0000002079097957__p1999514315198">LIBRARY_PATH</p>
</td>
<td class="cellrowborder" valign="top" width="34.505050505050505%" headers="mcps1.3.4.2.4.1.2 "><p id="EN-US_TOPIC_0000002079097957__p899563116194">Static library paths</p>
</td>
<td class="cellrowborder" valign="top" width="44.31313131313131%" headers="mcps1.3.4.2.4.1.3 "><p id="EN-US_TOPIC_0000002079097957__p149959313198">LIBRARY_PATH=/usr/local/cuda/lib64/stubs</p>
</td>
</tr>
<tr id="EN-US_TOPIC_0000002079097957__row0995133114195"><td class="cellrowborder" valign="top" width="21.181818181818183%" headers="mcps1.3.4.2.4.1.1 "><p id="EN-US_TOPIC_0000002079097957__p1099563110191">MA_HOME</p>
</td>
<td class="cellrowborder" valign="top" width="34.505050505050505%" headers="mcps1.3.4.2.4.1.2 "><p id="EN-US_TOPIC_0000002079097957__p2099553118194">Main directory of a training job</p>
</td>
<td class="cellrowborder" valign="top" width="44.31313131313131%" headers="mcps1.3.4.2.4.1.3 "><p id="EN-US_TOPIC_0000002079097957__p49951431131917">MA_HOME=/home/ma-user</p>
</td>
</tr>
<tr id="EN-US_TOPIC_0000002079097957__row3995103111913"><td class="cellrowborder" valign="top" width="21.181818181818183%" headers="mcps1.3.4.2.4.1.1 "><p id="EN-US_TOPIC_0000002079097957__p3995143131910">MA_JOB_DIR</p>
</td>
<td class="cellrowborder" valign="top" width="34.505050505050505%" headers="mcps1.3.4.2.4.1.2 "><p id="EN-US_TOPIC_0000002079097957__p199573181919">Parent directory of the training algorithm folder</p>
</td>
<td class="cellrowborder" valign="top" width="44.31313131313131%" headers="mcps1.3.4.2.4.1.3 "><p id="EN-US_TOPIC_0000002079097957__p99955314196">MA_JOB_DIR=/home/ma-user/modelarts/user-job-dir</p>
</td>
</tr>
<tr id="EN-US_TOPIC_0000002079097957__row39959315199"><td class="cellrowborder" valign="top" width="21.181818181818183%" headers="mcps1.3.4.2.4.1.1 "><p id="EN-US_TOPIC_0000002079097957__p19951431161916">MA_MOUNT_PATH</p>
</td>
<td class="cellrowborder" valign="top" width="34.505050505050505%" headers="mcps1.3.4.2.4.1.2 "><p id="EN-US_TOPIC_0000002079097957__p299513161912">Path mounted to a ModelArts training container, which is used to temporarily store training algorithms, algorithm input, algorithm output, and logs</p>
</td>
<td class="cellrowborder" valign="top" width="44.31313131313131%" headers="mcps1.3.4.2.4.1.3 "><p id="EN-US_TOPIC_0000002079097957__p16995123114192">MA_MOUNT_PATH=/home/ma-user/modelarts</p>
</td>
</tr>
<tr id="EN-US_TOPIC_0000002079097957__row2995143131910"><td class="cellrowborder" valign="top" width="21.181818181818183%" headers="mcps1.3.4.2.4.1.1 "><p id="EN-US_TOPIC_0000002079097957__p8995631111916">MA_LOG_DIR</p>
</td>
<td class="cellrowborder" valign="top" width="34.505050505050505%" headers="mcps1.3.4.2.4.1.2 "><p id="EN-US_TOPIC_0000002079097957__p2995631141912">Training log directory</p>
</td>
<td class="cellrowborder" valign="top" width="44.31313131313131%" headers="mcps1.3.4.2.4.1.3 "><p id="EN-US_TOPIC_0000002079097957__p10996193117198">MA_LOG_DIR=/home/ma-user/modelarts/log</p>
</td>
</tr>
<tr id="EN-US_TOPIC_0000002079097957__row1399623111196"><td class="cellrowborder" valign="top" width="21.181818181818183%" headers="mcps1.3.4.2.4.1.1 "><p id="EN-US_TOPIC_0000002079097957__p1299673111915">MA_SCRIPT_INTERPRETER</p>
</td>
<td class="cellrowborder" valign="top" width="34.505050505050505%" headers="mcps1.3.4.2.4.1.2 "><p id="EN-US_TOPIC_0000002079097957__p7996131141914">Training script interpreter</p>
</td>
<td class="cellrowborder" valign="top" width="44.31313131313131%" headers="mcps1.3.4.2.4.1.3 "><p id="EN-US_TOPIC_0000002079097957__p799653111913">MA_SCRIPT_INTERPRETER=</p>
</td>
</tr>
<tr id="EN-US_TOPIC_0000002079097957__row8996153161911"><td class="cellrowborder" valign="top" width="21.181818181818183%" headers="mcps1.3.4.2.4.1.1 "><p id="EN-US_TOPIC_0000002079097957__p159967317197">WORKSPACE</p>
</td>
<td class="cellrowborder" valign="top" width="34.505050505050505%" headers="mcps1.3.4.2.4.1.2 "><p id="EN-US_TOPIC_0000002079097957__p149961631151917">Training algorithm directory</p>
</td>
<td class="cellrowborder" valign="top" width="44.31313131313131%" headers="mcps1.3.4.2.4.1.3 "><p id="EN-US_TOPIC_0000002079097957__p17996123112192">WORKSPACE=/home/ma-user/modelarts/user-job-dir/code</p>
</td>
</tr>
</tbody>
</table>
</div>
<div class="tablenoborder"><table cellpadding="4" cellspacing="0" summary="" id="EN-US_TOPIC_0000002079097957__table28771431162113" frame="border" border="1" rules="all"><caption><b>Table 2 </b>Environment variables of a distributed training job</caption><thead align="left"><tr id="EN-US_TOPIC_0000002079097957__row2090453110217"><th align="left" class="cellrowborder" valign="top" width="21.70707070707071%" id="mcps1.3.5.2.4.1.1"><p id="EN-US_TOPIC_0000002079097957__p13904731122119">Variable</p>
</th>
<th align="left" class="cellrowborder" valign="top" width="34.25252525252526%" id="mcps1.3.5.2.4.1.2"><p id="EN-US_TOPIC_0000002079097957__p79051931162116">Description</p>
</th>
<th align="left" class="cellrowborder" valign="top" width="44.04040404040404%" id="mcps1.3.5.2.4.1.3"><p id="EN-US_TOPIC_0000002079097957__p6905193114217">Example</p>
</th>
</tr>
</thead>
<tbody><tr id="EN-US_TOPIC_0000002079097957__row990573172115"><td class="cellrowborder" valign="top" width="21.70707070707071%" headers="mcps1.3.5.2.4.1.1 "><p id="EN-US_TOPIC_0000002079097957__p12905193162114">MA_CURRENT_IP</p>
</td>
<td class="cellrowborder" valign="top" width="34.25252525252526%" headers="mcps1.3.5.2.4.1.2 "><p id="EN-US_TOPIC_0000002079097957__p11905183113219">IP address of the physical node on which a job container is running.</p>
</td>
<td class="cellrowborder" valign="top" width="44.04040404040404%" headers="mcps1.3.5.2.4.1.3 "><p id="EN-US_TOPIC_0000002079097957__p59051531122116">MA_CURRENT_IP=192.168.23.38</p>
</td>
</tr>
<tr id="EN-US_TOPIC_0000002079097957__row1490517315219"><td class="cellrowborder" valign="top" width="21.70707070707071%" headers="mcps1.3.5.2.4.1.1 "><p id="EN-US_TOPIC_0000002079097957__p12905113112110">MA_NUM_GPUS</p>
</td>
<td class="cellrowborder" valign="top" width="34.25252525252526%" headers="mcps1.3.5.2.4.1.2 "><p id="EN-US_TOPIC_0000002079097957__p9905831152118">Number of GPUs used by a job container.</p>
</td>
<td class="cellrowborder" valign="top" width="44.04040404040404%" headers="mcps1.3.5.2.4.1.3 "><p id="EN-US_TOPIC_0000002079097957__p16905631152115">MA_NUM_GPUS=8</p>
</td>
</tr>
<tr id="EN-US_TOPIC_0000002079097957__row17905931192113"><td class="cellrowborder" valign="top" width="21.70707070707071%" headers="mcps1.3.5.2.4.1.1 "><p id="EN-US_TOPIC_0000002079097957__p109052311216">MA_TASK_NAME</p>
</td>
<td class="cellrowborder" valign="top" width="34.25252525252526%" headers="mcps1.3.5.2.4.1.2 "><p id="EN-US_TOPIC_0000002079097957__p3905173192111">Name of a job container, for example:</p>
<ul id="EN-US_TOPIC_0000002079097957__ul61672054134011"><li id="EN-US_TOPIC_0000002079097957__li616719548404"><strong id="EN-US_TOPIC_0000002079097957__b183721922161518">worker</strong> in MindSpore and PyTorch.</li><li id="EN-US_TOPIC_0000002079097957__li21676545402"><strong id="EN-US_TOPIC_0000002079097957__b1093954411158">learner</strong> or <strong id="EN-US_TOPIC_0000002079097957__b126354720157">worker</strong> in reinforcement learning engines.</li><li id="EN-US_TOPIC_0000002079097957__li11167254134010"><strong id="EN-US_TOPIC_0000002079097957__b1570101320179">ps</strong> or <strong id="EN-US_TOPIC_0000002079097957__b9373101521713">worker</strong> in <strong id="EN-US_TOPIC_0000002079097957__b1594321071616">TensorFlow</strong>.</li></ul>
</td>
<td class="cellrowborder" valign="top" width="44.04040404040404%" headers="mcps1.3.5.2.4.1.3 "><p id="EN-US_TOPIC_0000002079097957__p14905031142116">MA_TASK_NAME=worker</p>
</td>
</tr>
<tr id="EN-US_TOPIC_0000002079097957__row129054316216"><td class="cellrowborder" valign="top" width="21.70707070707071%" headers="mcps1.3.5.2.4.1.1 "><p id="EN-US_TOPIC_0000002079097957__p11905831172116">MA_NUM_HOSTS</p>
</td>
<td class="cellrowborder" valign="top" width="34.25252525252526%" headers="mcps1.3.5.2.4.1.2 "><p id="EN-US_TOPIC_0000002079097957__p11905103192119">Compute nodes required for a training job.</p>
</td>
<td class="cellrowborder" valign="top" width="44.04040404040404%" headers="mcps1.3.5.2.4.1.3 "><p id="EN-US_TOPIC_0000002079097957__p990563142116">MA_NUM_HOSTS=4</p>
</td>
</tr>
<tr id="EN-US_TOPIC_0000002079097957__row1690573172117"><td class="cellrowborder" valign="top" width="21.70707070707071%" headers="mcps1.3.5.2.4.1.1 "><p id="EN-US_TOPIC_0000002079097957__p19905193172119">VC_TASK_INDEX</p>
</td>
<td class="cellrowborder" valign="top" width="34.25252525252526%" headers="mcps1.3.5.2.4.1.2 "><p id="EN-US_TOPIC_0000002079097957__p179051331142119">Sequence number of a job container for multi-node training. The value of the first container is <strong id="EN-US_TOPIC_0000002079097957__b1664911386116">0</strong>.</p>
</td>
<td class="cellrowborder" valign="top" width="44.04040404040404%" headers="mcps1.3.5.2.4.1.3 "><p id="EN-US_TOPIC_0000002079097957__p1490533115219">VC_TASK_INDEX=0</p>
</td>
</tr>
<tr id="EN-US_TOPIC_0000002079097957__row13905203132112"><td class="cellrowborder" valign="top" width="21.70707070707071%" headers="mcps1.3.5.2.4.1.1 "><p id="EN-US_TOPIC_0000002079097957__p169051331192114">VC_WORKER_NUM</p>
</td>
<td class="cellrowborder" valign="top" width="34.25252525252526%" headers="mcps1.3.5.2.4.1.2 "><p id="EN-US_TOPIC_0000002079097957__p490516312216">Compute nodes required for a training job.</p>
</td>
<td class="cellrowborder" valign="top" width="44.04040404040404%" headers="mcps1.3.5.2.4.1.3 "><p id="EN-US_TOPIC_0000002079097957__p17905173112111">VC_WORKER_NUM=4</p>
</td>
</tr>
<tr id="EN-US_TOPIC_0000002079097957__row1590593111219"><td class="cellrowborder" valign="top" width="21.70707070707071%" headers="mcps1.3.5.2.4.1.1 "><p id="EN-US_TOPIC_0000002079097957__p10905103112212">VC_WORKER_HOSTS</p>
</td>
<td class="cellrowborder" valign="top" width="34.25252525252526%" headers="mcps1.3.5.2.4.1.2 "><p id="EN-US_TOPIC_0000002079097957__p9905153115218">Domain name of each node for multi-node training. Use commas (,) to separate the domain names in sequence. You can obtain the IP address through domain name resolution.</p>
</td>
<td class="cellrowborder" valign="top" width="44.04040404040404%" headers="mcps1.3.5.2.4.1.3 "><p id="EN-US_TOPIC_0000002079097957__p11905103132114">VC_WORKER_HOSTS=modelarts-job-a0978141-1712-4f9b-8a83-000000000000-worker-0.modelarts-job-a0978141-1712-4f9b-8a83-000000000000,modelarts-job-a0978141-1712-4f9b-8a83-000000000000-worker-1.ob-a0978141-1712-4f9b-8a83-000000000000,modelarts-job-a0978141-1712-4f9b-8a83-000000000000-worker-2.modelarts-job-a0978141-1712-4f9b-8a83-000000000000,ob-a0978141-1712-4f9b-8a83-000000000000-worker-3.modelarts-job-a0978141-1712-4f9b-8a83-000000000000</p>
</td>
</tr>
</tbody>
</table>
</div>
<div class="tablenoborder"><table cellpadding="4" cellspacing="0" summary="" id="EN-US_TOPIC_0000002079097957__table970554921918" frame="border" border="1" rules="all"><caption><b>Table 3 </b>NCCL environment variables</caption><thead align="left"><tr id="EN-US_TOPIC_0000002079097957__row572415496192"><th align="left" class="cellrowborder" valign="top" width="22.161616161616163%" id="mcps1.3.6.2.4.1.1"><p id="EN-US_TOPIC_0000002079097957__p1072424920195">Variable</p>
</th>
<th align="left" class="cellrowborder" valign="top" width="34.151515151515156%" id="mcps1.3.6.2.4.1.2"><p id="EN-US_TOPIC_0000002079097957__p972454913191">Description</p>
</th>
<th align="left" class="cellrowborder" valign="top" width="43.686868686868685%" id="mcps1.3.6.2.4.1.3"><p id="EN-US_TOPIC_0000002079097957__p37240496199">Example</p>
</th>
</tr>
</thead>
<tbody><tr id="EN-US_TOPIC_0000002079097957__row1472484912194"><td class="cellrowborder" valign="top" width="22.161616161616163%" headers="mcps1.3.6.2.4.1.1 "><p id="EN-US_TOPIC_0000002079097957__p12724449131911">NCCL_VERSION</p>
</td>
<td class="cellrowborder" valign="top" width="34.151515151515156%" headers="mcps1.3.6.2.4.1.2 "><p id="EN-US_TOPIC_0000002079097957__p37241649171912">NCCL version</p>
</td>
<td class="cellrowborder" valign="top" width="43.686868686868685%" headers="mcps1.3.6.2.4.1.3 "><p id="EN-US_TOPIC_0000002079097957__p17241149161913">NCCL_VERSION=2.7.8</p>
</td>
</tr>
<tr id="EN-US_TOPIC_0000002079097957__row1572494910198"><td class="cellrowborder" valign="top" width="22.161616161616163%" headers="mcps1.3.6.2.4.1.1 "><p id="EN-US_TOPIC_0000002079097957__p2724154961918">NCCL_DEBUG</p>
</td>
<td class="cellrowborder" valign="top" width="34.151515151515156%" headers="mcps1.3.6.2.4.1.2 "><p id="EN-US_TOPIC_0000002079097957__p1621914112018">NCCL log level</p>
</td>
<td class="cellrowborder" valign="top" width="43.686868686868685%" headers="mcps1.3.6.2.4.1.3 "><p id="EN-US_TOPIC_0000002079097957__p2724249111917">NCCL_DEBUG=INFO</p>
</td>
</tr>
<tr id="EN-US_TOPIC_0000002079097957__row57241149181918"><td class="cellrowborder" valign="top" width="22.161616161616163%" headers="mcps1.3.6.2.4.1.1 "><p id="EN-US_TOPIC_0000002079097957__p87241149191910">NCCL_IB_HCA</p>
</td>
<td class="cellrowborder" valign="top" width="34.151515151515156%" headers="mcps1.3.6.2.4.1.2 "><p id="EN-US_TOPIC_0000002079097957__p0724649191917">InfiniBand NIC to use for communication</p>
</td>
<td class="cellrowborder" valign="top" width="43.686868686868685%" headers="mcps1.3.6.2.4.1.3 "><p id="EN-US_TOPIC_0000002079097957__p19724849161917">NCCL_IB_HCA=^mlx5_bond_0</p>
</td>
</tr>
<tr id="EN-US_TOPIC_0000002079097957__row17242049131917"><td class="cellrowborder" valign="top" width="22.161616161616163%" headers="mcps1.3.6.2.4.1.1 "><p id="EN-US_TOPIC_0000002079097957__p6724164951920">NCCL_SOCKET_IFNAME</p>
</td>
<td class="cellrowborder" valign="top" width="34.151515151515156%" headers="mcps1.3.6.2.4.1.2 "><p id="EN-US_TOPIC_0000002079097957__p16724134912193">IP interface to use for communication</p>
</td>
<td class="cellrowborder" valign="top" width="43.686868686868685%" headers="mcps1.3.6.2.4.1.3 "><p id="EN-US_TOPIC_0000002079097957__p127241249191910">NCCL_SOCKET_IFNAME=bond0,eth0</p>
</td>
</tr>
</tbody>
</table>
</div>
<div class="tablenoborder"><table cellpadding="4" cellspacing="0" summary="" id="EN-US_TOPIC_0000002079097957__table167761732182014" frame="border" border="1" rules="all"><caption><b>Table 4 </b>OBS environment variables</caption><thead align="left"><tr id="EN-US_TOPIC_0000002079097957__row1479315320205"><th align="left" class="cellrowborder" valign="top" width="22.6969696969697%" id="mcps1.3.7.2.4.1.1"><p id="EN-US_TOPIC_0000002079097957__p1779310325206">Variable</p>
</th>
<th align="left" class="cellrowborder" valign="top" width="34.323232323232325%" id="mcps1.3.7.2.4.1.2"><p id="EN-US_TOPIC_0000002079097957__p479313323207">Description</p>
</th>
<th align="left" class="cellrowborder" valign="top" width="42.97979797979798%" id="mcps1.3.7.2.4.1.3"><p id="EN-US_TOPIC_0000002079097957__p1979303252018">Example</p>
</th>
</tr>
</thead>
<tbody><tr id="EN-US_TOPIC_0000002079097957__row7793173292012"><td class="cellrowborder" valign="top" width="22.6969696969697%" headers="mcps1.3.7.2.4.1.1 "><p id="EN-US_TOPIC_0000002079097957__p1379316322201">S3_ENDPOINT</p>
</td>
<td class="cellrowborder" valign="top" width="34.323232323232325%" headers="mcps1.3.7.2.4.1.2 "><p id="EN-US_TOPIC_0000002079097957__p17793123262012">OBS endpoint</p>
</td>
<td class="cellrowborder" valign="top" width="42.97979797979798%" headers="mcps1.3.7.2.4.1.3 "><p id="EN-US_TOPIC_0000002079097957__p014333435515">S3_ENDPOINT=https://obs.region.xxx.com</p>
</td>
</tr>
<tr id="EN-US_TOPIC_0000002079097957__row1279316322209"><td class="cellrowborder" valign="top" width="22.6969696969697%" headers="mcps1.3.7.2.4.1.1 "><p id="EN-US_TOPIC_0000002079097957__p1779333218203">S3_VERIFY_SSL</p>
</td>
<td class="cellrowborder" valign="top" width="34.323232323232325%" headers="mcps1.3.7.2.4.1.2 "><p id="EN-US_TOPIC_0000002079097957__p13793932112019">Whether to use SSL to access OBS</p>
</td>
<td class="cellrowborder" valign="top" width="42.97979797979798%" headers="mcps1.3.7.2.4.1.3 "><p id="EN-US_TOPIC_0000002079097957__p87931032162012">S3_VERIFY_SSL=0</p>
</td>
</tr>
<tr id="EN-US_TOPIC_0000002079097957__row12793153212015"><td class="cellrowborder" valign="top" width="22.6969696969697%" headers="mcps1.3.7.2.4.1.1 "><p id="EN-US_TOPIC_0000002079097957__p14793183214200">S3_USE_HTTPS</p>
</td>
<td class="cellrowborder" valign="top" width="34.323232323232325%" headers="mcps1.3.7.2.4.1.2 "><p id="EN-US_TOPIC_0000002079097957__p479333292012">Whether to use HTTPS to access OBS</p>
</td>
<td class="cellrowborder" valign="top" width="42.97979797979798%" headers="mcps1.3.7.2.4.1.3 "><p id="EN-US_TOPIC_0000002079097957__p9793132142010">S3_USE_HTTPS=1</p>
</td>
</tr>
</tbody>
</table>
</div>
<div class="tablenoborder"><table cellpadding="4" cellspacing="0" summary="" id="EN-US_TOPIC_0000002079097957__table1378715112216" frame="border" border="1" rules="all"><caption><b>Table 5 </b>Environment variables of the pip source and API Gateway address</caption><thead align="left"><tr id="EN-US_TOPIC_0000002079097957__row16781615122219"><th align="left" class="cellrowborder" valign="top" width="22.96969696969697%" id="mcps1.3.8.2.4.1.1"><p id="EN-US_TOPIC_0000002079097957__p18781515112219">Variable</p>
</th>
<th align="left" class="cellrowborder" valign="top" width="33.88888888888889%" id="mcps1.3.8.2.4.1.2"><p id="EN-US_TOPIC_0000002079097957__p078115192211">Description</p>
</th>
<th align="left" class="cellrowborder" valign="top" width="43.14141414141414%" id="mcps1.3.8.2.4.1.3"><p id="EN-US_TOPIC_0000002079097957__p197861511229">Example</p>
</th>
</tr>
</thead>
<tbody><tr id="EN-US_TOPIC_0000002079097957__row479101510224"><td class="cellrowborder" valign="top" width="22.96969696969697%" headers="mcps1.3.8.2.4.1.1 "><p id="EN-US_TOPIC_0000002079097957__p20793155223">MA_PIP_HOST</p>
</td>
<td class="cellrowborder" valign="top" width="33.88888888888889%" headers="mcps1.3.8.2.4.1.2 "><p id="EN-US_TOPIC_0000002079097957__p17931512212">Domain name of the pip source</p>
</td>
<td class="cellrowborder" valign="top" width="43.14141414141414%" headers="mcps1.3.8.2.4.1.3 "><p id="EN-US_TOPIC_0000002079097957__p10791315192217">MA_PIP_HOST=repo.xxx.com</p>
</td>
</tr>
<tr id="EN-US_TOPIC_0000002079097957__row27918151224"><td class="cellrowborder" valign="top" width="22.96969696969697%" headers="mcps1.3.8.2.4.1.1 "><p id="EN-US_TOPIC_0000002079097957__p07911520225">MA_PIP_URL</p>
</td>
<td class="cellrowborder" valign="top" width="33.88888888888889%" headers="mcps1.3.8.2.4.1.2 "><p id="EN-US_TOPIC_0000002079097957__p079101520224">Address of the pip source</p>
</td>
<td class="cellrowborder" valign="top" width="43.14141414141414%" headers="mcps1.3.8.2.4.1.3 "><p id="EN-US_TOPIC_0000002079097957__p127951552211">MA_PIP_URL=http://repo.xxx.com/repository/pypi/simple/</p>
</td>
</tr>
<tr id="EN-US_TOPIC_0000002079097957__row579121532213"><td class="cellrowborder" valign="top" width="22.96969696969697%" headers="mcps1.3.8.2.4.1.1 "><p id="EN-US_TOPIC_0000002079097957__p9794154226">MA_APIGW_ENDPOINT</p>
</td>
<td class="cellrowborder" valign="top" width="33.88888888888889%" headers="mcps1.3.8.2.4.1.2 "><p id="EN-US_TOPIC_0000002079097957__p679115122214">ModelArts API Gateway address</p>
</td>
<td class="cellrowborder" valign="top" width="43.14141414141414%" headers="mcps1.3.8.2.4.1.3 "><p id="EN-US_TOPIC_0000002079097957__p57901582218">MA_APIGW_ENDPOINT=https://modelarts.region.xxx.xxx.com</p>
</td>
</tr>
</tbody>
</table>
</div>
<div class="tablenoborder"><table cellpadding="4" cellspacing="0" summary="" id="EN-US_TOPIC_0000002079097957__table1859236102213" frame="border" border="1" rules="all"><caption><b>Table 6 </b>Environment variables of job metadata </caption><thead align="left"><tr id="EN-US_TOPIC_0000002079097957__row76111869223"><th align="left" class="cellrowborder" valign="top" width="22.96969696969697%" id="mcps1.3.9.2.4.1.1"><p id="EN-US_TOPIC_0000002079097957__p1611166132218">Variable</p>
</th>
<th align="left" class="cellrowborder" valign="top" width="34.54545454545455%" id="mcps1.3.9.2.4.1.2"><p id="EN-US_TOPIC_0000002079097957__p46111672215">Description</p>
</th>
<th align="left" class="cellrowborder" valign="top" width="42.484848484848484%" id="mcps1.3.9.2.4.1.3"><p id="EN-US_TOPIC_0000002079097957__p7611136182210">Example</p>
</th>
</tr>
</thead>
<tbody><tr id="EN-US_TOPIC_0000002079097957__row861166102216"><td class="cellrowborder" valign="top" width="22.96969696969697%" headers="mcps1.3.9.2.4.1.1 "><p id="EN-US_TOPIC_0000002079097957__p1361111618224">MA_CURRENT_INSTANCE_NAME</p>
</td>
<td class="cellrowborder" valign="top" width="34.54545454545455%" headers="mcps1.3.9.2.4.1.2 "><p id="EN-US_TOPIC_0000002079097957__p1461113602220">Name of the current node for multi-node training</p>
</td>
<td class="cellrowborder" valign="top" width="42.484848484848484%" headers="mcps1.3.9.2.4.1.3 "><p id="EN-US_TOPIC_0000002079097957__p1761196172219">MA_CURRENT_INSTANCE_NAME=modelarts-job-a0978141-1712-4f9b-8a83-000000000000-worker-1</p>
</td>
</tr>
</tbody>
</table>
</div>
<div class="tablenoborder"><table cellpadding="4" cellspacing="0" summary="" id="EN-US_TOPIC_0000002079097957__table4530451145018" frame="border" border="1" rules="all"><caption><b>Table 7 </b>Precheck environment variables</caption><thead align="left"><tr id="EN-US_TOPIC_0000002079097957__row65306517505"><th align="left" class="cellrowborder" valign="top" width="22.96969696969697%" id="mcps1.3.10.2.4.1.1"><p id="EN-US_TOPIC_0000002079097957__p12530175115012">Variable</p>
</th>
<th align="left" class="cellrowborder" valign="top" width="34.54545454545455%" id="mcps1.3.10.2.4.1.2"><p id="EN-US_TOPIC_0000002079097957__p75308519502">Description</p>
</th>
<th align="left" class="cellrowborder" valign="top" width="42.484848484848484%" id="mcps1.3.10.2.4.1.3"><p id="EN-US_TOPIC_0000002079097957__p0530951135014">Example</p>
</th>
</tr>
</thead>
<tbody><tr id="EN-US_TOPIC_0000002079097957__row95304518504"><td class="cellrowborder" valign="top" width="22.96969696969697%" headers="mcps1.3.10.2.4.1.1 "><p id="EN-US_TOPIC_0000002079097957__p15850121510517">MA_DETECT_TRAIN_INJECT_CODE</p>
</td>
<td class="cellrowborder" valign="top" width="34.54545454545455%" headers="mcps1.3.10.2.4.1.2 "><p id="EN-US_TOPIC_0000002079097957__p11396122205115">Whether to enable ModelArts precheck.</p>
<p id="EN-US_TOPIC_0000002079097957__p12396162215519">The default value is <strong id="EN-US_TOPIC_0000002079097957__b329411369516">1</strong>, indicating that precheck is enabled.</p>
<p id="EN-US_TOPIC_0000002079097957__p113961922195116">The value <strong id="EN-US_TOPIC_0000002079097957__b111319408519">0</strong> indicates that the precheck is disabled.</p>
<p id="EN-US_TOPIC_0000002079097957__p13961522185112">Enable precheck to detect node and driver faults before they affect services.</p>
</td>
<td class="cellrowborder" valign="top" width="42.484848484848484%" headers="mcps1.3.10.2.4.1.3 "><p id="EN-US_TOPIC_0000002079097957__p15530175175013">1</p>
</td>
</tr>
</tbody>
</table>
</div>
</div>
<div>
<div class="familylinks">
<div class="parentlink"><strong>Parent topic:</strong> <a href="develop-modelarts-0010.html">Performing a Training</a></div>
</div>
</div>
<script language="JavaScript">
<!--
image_size('.imgResize');
var msg_imageMax = "view original image";
var msg_imageClose = "close";
//--></script>
View Git Blame Copy Permalink