Open SamuelKocur opened 1 month ago
It is because, the agent decides whether this iteration gives a retrieved information through which sub-question can be answered or not ...
If not, instead of framing the retrieved information to final response, it moves to next iteration. This way, if it ends up seeing awful information after each retrieval iteration... Then it stops the process, apologizing for not being able to find relevant information
.
Hope this helps.
In documenation is written: Max Iter is the maximum number of iterations the agent can perform before being forced to give its best answer.
Why does the agent return: "Agent stopped due to iteration limit or time limit."? Shouldn't it rather try to generate answer based on the context it gathered?