[Feature Request]: Support for Asynchronous Processing in llms.watsonx

[mdielacher]

Feature Description

Currently, llms.watsonx does not support asynchronous processing, which is crucial for the efficient operation of evaluators such as the FaithfulnessEvaluator and RelevancyEvaluator. In comparison llms.ollama/llms.openai do support asynchronous processing, enabling smoother and faster evaluation within pipelines.

Reason

The classes:

• acomplete
• astream_chat
• achat
• astream_complete

currently marked as NotImplemented for llms.watsonx.

Value of Feature

These evaluators are essential for ensuring the quality of responses.

[beko-dt]

+1 Would be valuable to have this!