You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Great to see one of the fundamental concepts to agents being implemented.
At the moment is seems that every agent in the crew is provided with the knowledge - are you planning on enabling the scope reduction to certain agents or where multiple knowledge articles are loaded which agents have access to which knowledge artifacts? Or should we design it so all agents in the crew will have access to the knowledge so therefore splitting up into multiple crews?
Obviously providing knowledge scope to an agent reduces the amount of tokens processed for every agent especially if certain agents don't need this knowledge, or where multiple knowledge artifacts are present and certain agents only need to know about a subset of this knowledge.
Also i would be useful if the metadata was provided to the agent (this seems to be missing), so that we could suggest in the task description to base their output on this as at the moment this is missing from the prompt that is generated.
I guess it would also be useful to be able to use the output of the "crew.trian" process, so the .pkl file can act as one of the knowledge artifacts provided to the crew with any other knowledge artifacts when you run it in the future.
reacted with thumbs up emoji reacted with thumbs down emoji reacted with laugh emoji reacted with hooray emoji reacted with confused emoji reacted with heart emoji reacted with rocket emoji reacted with eyes emoji
-
Great to see one of the fundamental concepts to agents being implemented.
At the moment is seems that every agent in the crew is provided with the knowledge - are you planning on enabling the scope reduction to certain agents or where multiple knowledge articles are loaded which agents have access to which knowledge artifacts? Or should we design it so all agents in the crew will have access to the knowledge so therefore splitting up into multiple crews?
Obviously providing knowledge scope to an agent reduces the amount of tokens processed for every agent especially if certain agents don't need this knowledge, or where multiple knowledge artifacts are present and certain agents only need to know about a subset of this knowledge.
Also i would be useful if the metadata was provided to the agent (this seems to be missing), so that we could suggest in the task description to base their output on this as at the moment this is missing from the prompt that is generated.
I guess it would also be useful to be able to use the output of the "crew.trian" process, so the .pkl file can act as one of the knowledge artifacts provided to the crew with any other knowledge artifacts when you run it in the future.
Beta Was this translation helpful? Give feedback.
All reactions