NEW Fine-Tuning OpenAI Models: A Guide 🚀

Document Retrieval

If you want to start training document retrieval or recommender models, you might want to develop a dataset with that identifies similar documents. Use this template to identify and choose documents that are related to a specific query or an existing document.

Interactive Template Preview

Labeling Configuration

<View>
  <Text name="query" value="$query" />
  <Header value="Select document related to the query:" />
  <View style="display: flex">
    <View>
    <Text name="text1" value="$text1" />
    <Text name="text2" value="$text2" />
    <Text name="text3" value="$text3" />
    </View>
    <View style="padding: 30px">
    <Choices name="selection" toName="query" required="true" choice="multiple">
      <Choice value="One" />
      <Choice value="Two" />
      <Choice value="Three" />
    </Choices>
    </View>
  </View>
</View>

About the labeling configuration

All labeling configurations must be wrapped in View tags.

Use the Text object tag to specify the text data to be annotated:

<Text name="query" value="$query" />

You can add a header to provide instructions to the annotator:

<Header value="Select document related to the query:" />
```  

Add styling to the [View](/tags/view.html) tag to control the appearance of the text samples and choices
```xml
<View style="display: flex">

Wrap the text snippets in a new View tag and use the Text object tag to display 3 other text snippets:

<View>
    <Text name="text1" value="$text1" />
    <Text name="text2" value="$text2" />
    <Text name="text3" value="$text3" />
</View>

Add styling to another View tag to place the choices next to the text samples, then use the Choices control tag to require a selection from annotators and allow them to select multiple text snippets that apply:

<View style="padding: 30px">
    <Choices name="selection" toName="query" required="true" choice="multiple">
      <Choice value="One" />
      <Choice value="Two" />
      <Choice value="Three" />
    </Choices>
</View>