Which feature of Language Studio does not require a model to be trained?

The feature of Language Studio that does not require a model to be trained is Analyze sentiment. This feature leverages pre-trained models that are already available within the Azure ecosystem. When you use Analyze sentiment, you can quickly and easily assess the emotional tone behind a body of text without needing to spend time training a model specifically for your data. This is particularly useful for organizations looking to implement sentiment analysis rapidly, as they can utilize the existing capabilities of Azure without the need for additional machine learning development efforts.

In contrast, other features such as Conversational language understanding, Custom text classification, and Language translation typically involve either custom model training or configuration to align with specific use cases or datasets. For instance, conversational language understanding usually requires training for particular intents and entities relevant to the application's context. Similarly, custom text classification often necessitates building a model tailored to the specific types of text and classification needs that differ from what general models might provide. Language translation can also require fine-tuning to achieve high accuracy, especially when dealing with specialized terminology or particular dialects. Thus, the pre-trained nature of Analyze sentiment distinguishes it from these other options.