doc-exports/docs/modelarts/umn/datalabel-modelarts_0016.html
Lai, Weijian 4e4b2d5f6d ModelArts UMN 23.3.0 Version.
Reviewed-by: Hasko, Vladimir <vladimir.hasko@t-systems.com>
Co-authored-by: Lai, Weijian <laiweijian4@huawei.com>
Co-committed-by: Lai, Weijian <laiweijian4@huawei.com>
2024-06-26 07:03:02 +00:00

42 lines
14 KiB
HTML

<a name="EN-US_TOPIC_0000001910027666"></a><a name="EN-US_TOPIC_0000001910027666"></a>
<h1 class="topictitle1">Speech Paragraph Labeling</h1>
<div id="body8662426"><p id="EN-US_TOPIC_0000001910027666__en-us_topic_0000001185384421_p1955228172214">Model training requires a large amount of labeled data. Therefore, before the model training, label the unlabeled audio files. ModelArts enables you to label audio files. In addition, you can modify the labels of audio files, or remove their labels and label the audio files again.</p>
<div class="section" id="EN-US_TOPIC_0000001910027666__en-us_topic_0000001185384421_section139520290612"><h4 class="sectiontitle">Starting Labeling</h4><ol id="EN-US_TOPIC_0000001910027666__en-us_topic_0000001185384421_en-us_topic_0000001185384417_ol1332113431875"><li id="EN-US_TOPIC_0000001910027666__en-us_topic_0000001185384421_en-us_topic_0000001185384417_li173221243078">Log in to the ModelArts management console. In the navigation pane on the left, choose <strong id="EN-US_TOPIC_0000001910027666__b53136059942726">Data Management</strong> &gt; <span class="parmname" id="EN-US_TOPIC_0000001910027666__parmname42571388542726"><b>Label Data</b></span>.</li><li id="EN-US_TOPIC_0000001910027666__en-us_topic_0000001185384421_en-us_topic_0000001185384417_li1241123818438">In the labeling job list, select a labeling type from the <strong id="EN-US_TOPIC_0000001910027666__b99725321914">All type</strong> drop-down list, click the job to be performed based on the labeling type. The details page of the job is displayed.</li><li id="EN-US_TOPIC_0000001910027666__en-us_topic_0000001185384421_en-us_topic_0000001185384417_li1766010993710">The job details page displays all data of the labeling job.</li></ol>
</div>
<div class="section" id="EN-US_TOPIC_0000001910027666__en-us_topic_0000001185384421_section888013232618"><h4 class="sectiontitle">Synchronizing Data Sources</h4><p id="EN-US_TOPIC_0000001910027666__en-us_topic_0000001185384421_p529632915615">ModelArts automatically synchronizes data and labeling information from datasets to the labeling job.</p>
<p id="EN-US_TOPIC_0000001910027666__en-us_topic_0000001185384421_p1029692918617">To quickly obtain the latest data in the OBS bucket, click <span class="parmname" id="EN-US_TOPIC_0000001910027666__en-us_topic_0000001185384421_parmname18209217452"><b>Synchronize Data Source</b></span> in the <span class="parmname" id="EN-US_TOPIC_0000001910027666__en-us_topic_0000001185384421_parmname92186171154"><b>Unlabeled</b></span> tab of the labeling job details page to add the data uploaded using OBS to the dataset.</p>
</div>
<div class="section" id="EN-US_TOPIC_0000001910027666__en-us_topic_0000001185384421_section888019266174"><h4 class="sectiontitle">Labeling Audio Files</h4><p id="EN-US_TOPIC_0000001910027666__en-us_topic_0000001185384421_p193441038156">The labeling job details page displays the <strong id="EN-US_TOPIC_0000001910027666__en-us_topic_0000001185384421_b4358151181418">Unlabeled</strong> and <strong id="EN-US_TOPIC_0000001910027666__en-us_topic_0000001185384421_b43586161413">Labeled</strong> tabs. The <strong id="EN-US_TOPIC_0000001910027666__en-us_topic_0000001185384421_b18358101131410">Unlabeled</strong> tab is displayed by default.</p>
<ol id="EN-US_TOPIC_0000001910027666__en-us_topic_0000001185384421_ol2907135216244"><li id="EN-US_TOPIC_0000001910027666__en-us_topic_0000001185384421_li127526819118">In the audio file list in the <strong id="EN-US_TOPIC_0000001910027666__en-us_topic_0000001185384421_b4795134481418">Unlabeled</strong> tab, click the target audio file. In the area on the right, the audio file is displayed. Click <span><img id="EN-US_TOPIC_0000001910027666__en-us_topic_0000001185384421_image19405165811512" src="figure/en-us_image_0000001910067874.png"></span> below the audio file to play the audio.</li><li id="EN-US_TOPIC_0000001910027666__en-us_topic_0000001185384421_li7639144112718">Select an audio segment based on the content being played, and enter the audio file label and content in the <span class="parmname" id="EN-US_TOPIC_0000001910027666__en-us_topic_0000001185384421_parmname722035313518"><b>Speech Content</b></span> text box.</li><li id="EN-US_TOPIC_0000001910027666__en-us_topic_0000001185384421_li122591050672">After entering the content, click <strong id="EN-US_TOPIC_0000001910027666__en-us_topic_0000001185384421_b13564820151819">Label</strong> to complete the labeling. The audio file is automatically moved to the <span class="wintitle" id="EN-US_TOPIC_0000001910027666__en-us_topic_0000001185384421_wintitle52181825141818"><b>Labeled</b></span> tab.</li></ol>
</div>
<div class="section" id="EN-US_TOPIC_0000001910027666__en-us_topic_0000001185384421_section14388174005916"><h4 class="sectiontitle">Viewing the Labeled Audio Files</h4><p id="EN-US_TOPIC_0000001910027666__en-us_topic_0000001185384421_p138711071810">On the labeling job details page, click the <strong id="EN-US_TOPIC_0000001910027666__en-us_topic_0000001185384421_b106390299187">Labeled</strong> tab to view the list of labeled audio files. Click the audio file to view the labeling information on the right.</p>
</div>
<div class="section" id="EN-US_TOPIC_0000001910027666__en-us_topic_0000001185384421_section0534612151819"><h4 class="sectiontitle">Modifying Labeled Data</h4><p id="EN-US_TOPIC_0000001910027666__en-us_topic_0000001185384421_p1981864110595">After labeling data, you can modify labeled data in the <span class="wintitle" id="EN-US_TOPIC_0000001910027666__wintitle1224564075117"><b>Labeled</b></span> tab.</p>
<ul id="EN-US_TOPIC_0000001910027666__en-us_topic_0000001185384421_ul1943393102017"><li id="EN-US_TOPIC_0000001910027666__en-us_topic_0000001185384421_li1946992912019">Modifying a label: On the labeling details page, click the <strong id="EN-US_TOPIC_0000001910027666__b15496144715233">Labeled</strong> tab, and select the audio file to be modified from the audio file list. In the right area, modify labeling information and click <strong id="EN-US_TOPIC_0000001910027666__b670763132416">Label</strong> to complete the modification.</li><li id="EN-US_TOPIC_0000001910027666__en-us_topic_0000001185384421_li1147062914207">Deleting a label: Click <strong id="EN-US_TOPIC_0000001910027666__b1173455510248">Delete</strong> in the <strong id="EN-US_TOPIC_0000001910027666__b15734455182410">Operation</strong> column of the target number to delete the label of the audio segment. Alternatively, you can click <span><img id="EN-US_TOPIC_0000001910027666__image825324319567" src="figure/en-us_image_0000001943987085.png"></span> above the labeled audio file to delete the label. Then click <strong id="EN-US_TOPIC_0000001910027666__b87131913122514">Label</strong>.</li></ul>
</div>
<div class="section" id="EN-US_TOPIC_0000001910027666__en-us_topic_0000001185384421_section15984542128"><h4 class="sectiontitle">Adding an Audio File</h4><p id="EN-US_TOPIC_0000001910027666__en-us_topic_0000001185384421_en-us_topic_0170889731_p266117351147">In addition to the data synchronized, you can directly add data on labeling job details page for labeling.</p>
<ol id="EN-US_TOPIC_0000001910027666__en-us_topic_0000001185384421_en-us_topic_0170889731_ol429210266513"><li id="EN-US_TOPIC_0000001910027666__en-us_topic_0000001185384421_en-us_topic_0170889731_li1924993814520">On the labeling job details page, click the <strong id="EN-US_TOPIC_0000001910027666__en-us_topic_0000001185384421_b727651343215">Unlabeled</strong> tab, click <strong id="EN-US_TOPIC_0000001910027666__en-us_topic_0000001185384421_b82766139321">Add data</strong> in the upper left corner.</li><li id="EN-US_TOPIC_0000001910027666__en-us_topic_0000001185384421_li01454543219">Configure input data and click <span class="uicontrol" id="EN-US_TOPIC_0000001910027666__en-us_topic_0000001185384421_uicontrol13260132123217"><b>OK</b></span>.<p id="EN-US_TOPIC_0000001910027666__en-us_topic_0000001185384421_p2070881211273"></p>
</li></ol>
</div>
<div class="section" id="EN-US_TOPIC_0000001910027666__en-us_topic_0000001185384421_section205685916013"><h4 class="sectiontitle">Deleting Audio Files</h4><p id="EN-US_TOPIC_0000001910027666__en-us_topic_0000001185384421_p3752212171611">You can quickly delete the audio files you want to discard.</p>
<p id="EN-US_TOPIC_0000001910027666__en-us_topic_0000001185384421_p1614812015518">In the <span class="wintitle" id="EN-US_TOPIC_0000001910027666__wintitle1782885925110"><b>Unlabeled</b></span> or <span class="wintitle" id="EN-US_TOPIC_0000001910027666__wintitle48299591514"><b>Labeled</b></span> tab, select the audio files to be deleted, and then click <span class="uicontrol" id="EN-US_TOPIC_0000001910027666__uicontrol13829185925119"><b>Delete File</b></span> in the upper left corner. In the displayed dialog box, select or deselect <span class="parmname" id="EN-US_TOPIC_0000001910027666__parmname1282965975114"><b>Delete the source files from OBS</b></span> as required. After confirmation, click <span class="uicontrol" id="EN-US_TOPIC_0000001910027666__uicontrol18829259115113"><b>OK</b></span> to delete the audio files.</p>
<div class="note" id="EN-US_TOPIC_0000001910027666__en-us_topic_0000001185384421_note10831343207"><img src="public_sys-resources/note_3.0-en-us.png"><span class="notetitle"> </span><div class="notebody"><p id="EN-US_TOPIC_0000001910027666__en-us_topic_0000001185384421_p17833431906">If you select <span class="parmname" id="EN-US_TOPIC_0000001910027666__parmname19211318624"><b>Delete the source files from OBS</b></span>, audio files stored in the corresponding OBS directory will be deleted when you delete the selected audio files. Deleting source files may affect other dataset versions or datasets using those files. As a result, the page display, training, or inference is abnormal. Deleted data cannot be recovered. Exercise caution when performing this operation.</p>
</div></div>
</div>
<div class="section" id="EN-US_TOPIC_0000001910027666__en-us_topic_0000001185384421_section11766134133017"><h4 class="sectiontitle">Managing Annotators</h4><p id="EN-US_TOPIC_0000001910027666__en-us_topic_0000001398306840_en-us_topic_0000001185384417_p0400144621811">If team labeling is enabled for a labeling job, view its labeling details in the <strong id="EN-US_TOPIC_0000001910027666__b189569565216">Annotator Management</strong> tab. Additionally, you can add, modify, or delete annotators.</p>
<ol id="EN-US_TOPIC_0000001910027666__en-us_topic_0000001398306840_ol14491310172315"><li id="EN-US_TOPIC_0000001910027666__en-us_topic_0000001398306840_li049171011232">Choose <strong id="EN-US_TOPIC_0000001910027666__b114541035203416">Data Management</strong> &gt; <strong id="EN-US_TOPIC_0000001910027666__b19454535173420">Label Data</strong>. In the <strong id="EN-US_TOPIC_0000001910027666__b1445413352340">My Creations</strong> or <strong id="EN-US_TOPIC_0000001910027666__b1045515359348">My Participations</strong> tab, view the list of all labeling jobs.</li><li id="EN-US_TOPIC_0000001910027666__en-us_topic_0000001398306840_li13948142542315">Locate the row that contains the target team labeling job. (The name of a team labeling job is followed by <span><img id="EN-US_TOPIC_0000001910027666__image2628737103415" src="figure/en-us_image_0000001910027806.png"></span>.)</li><li id="EN-US_TOPIC_0000001910027666__en-us_topic_0000001398306840_li176669516250">Choose <strong id="EN-US_TOPIC_0000001910027666__b48453100994851">More</strong> &gt; <strong id="EN-US_TOPIC_0000001910027666__b141160259794851">Annotator Management</strong> in the <strong id="EN-US_TOPIC_0000001910027666__b93269455494851">Operation</strong> column. Alternatively, click the job name to go to the job details page, and choose <strong id="EN-US_TOPIC_0000001910027666__b13520427794851">Team Labeling</strong> &gt; <strong id="EN-US_TOPIC_0000001910027666__b114156827394851">Annotator Management</strong> in the upper right corner.</li></ol>
<ul id="EN-US_TOPIC_0000001910027666__en-us_topic_0000001398306840_en-us_topic_0000001185384417_ul143722450281"><li id="EN-US_TOPIC_0000001910027666__en-us_topic_0000001398306840_en-us_topic_0000001185384417_li537294582812">Adding an annotator<p id="EN-US_TOPIC_0000001910027666__en-us_topic_0000001398306840_en-us_topic_0000001185384417_p723845392118"><a name="EN-US_TOPIC_0000001910027666__en-us_topic_0000001398306840_en-us_topic_0000001185384417_li537294582812"></a><a name="en-us_topic_0000001398306840_en-us_topic_0000001185384417_li537294582812"></a>Click <strong id="EN-US_TOPIC_0000001910027666__b59508956094851">Add Member</strong>, select a member name, and click <strong id="EN-US_TOPIC_0000001910027666__b120751623694851">OK</strong>.</p>
<p id="EN-US_TOPIC_0000001910027666__en-us_topic_0000001398306840_en-us_topic_0000001185384417_p177231347112320">Click <strong id="EN-US_TOPIC_0000001910027666__b39776963494851">Send Email</strong> in the <strong id="EN-US_TOPIC_0000001910027666__b19049256594851">Operation</strong> column to send the labeling job to the annotator by email.</p>
</li><li id="EN-US_TOPIC_0000001910027666__en-us_topic_0000001398306840_en-us_topic_0000001185384417_li261614352915">Modifying annotator information<p id="EN-US_TOPIC_0000001910027666__en-us_topic_0000001398306840_en-us_topic_0000001185384417_p744716392518"><a name="EN-US_TOPIC_0000001910027666__en-us_topic_0000001398306840_en-us_topic_0000001185384417_li261614352915"></a><a name="en-us_topic_0000001398306840_en-us_topic_0000001185384417_li261614352915"></a>Click <strong id="EN-US_TOPIC_0000001910027666__b121564333594851">Modify</strong> in the <strong id="EN-US_TOPIC_0000001910027666__b104112627794851">Operation</strong> column to modify the role of the annotator.</p>
</li><li id="EN-US_TOPIC_0000001910027666__en-us_topic_0000001398306840_en-us_topic_0000001185384417_li57521789299">Deleting an annotator<p id="EN-US_TOPIC_0000001910027666__en-us_topic_0000001398306840_en-us_topic_0000001185384417_p16917111112269"><a name="EN-US_TOPIC_0000001910027666__en-us_topic_0000001398306840_en-us_topic_0000001185384417_li57521789299"></a><a name="en-us_topic_0000001398306840_en-us_topic_0000001185384417_li57521789299"></a>Click <strong id="EN-US_TOPIC_0000001910027666__b3982508494851">Delete</strong> in the <strong id="EN-US_TOPIC_0000001910027666__b146759943594851">Operation</strong> column to delete the annotator.</p>
</li></ul>
</div>
</div>
<div>
<div class="familylinks">
<div class="parentlink"><strong>Parent topic:</strong> <a href="datalabel-modelarts_0013.html">Audio Labeling</a></div>
</div>
</div>