Using the watsonx.ai account provided with the Content Assistant
The Content Assistant, by default, uses a specific watsonx.ai account.
This default watsonx.ai account that is provided with the Content Assistant has the following capabilities:
- The ability to use the watsonx.ai hosted large language models to generate AI-based responses and summaries. For more information, see the topic Granite foundation models.
- The ability to use the watsonx.ai hosted embedding models to generate embeddings, which are used to generate responses. For more information, see the topic Slate foundation models.
License-related restrictions with the watsonx.ai account provided with Content Assistant
The following restrictions apply when you use the watsonx.ai account that is provided with the
Content Assistant:
- Limited choice of LLMs - You are restricted to use the current version of the IBM Granite model as the LLM.
- Limited choice of embedding models - You are restricted to use the current version of the IBM slate model as the embedding model.
- When you are using the default watsonx.ai account, you may not increase the default number of GenAI queue sweep worker threads. The default value is one worker thread per Content Platform Engine server. This limitation impacts both the throughput for documentation vector indexing and the throughput for document summarization.