The "use_docker" arg in an agent's code_execution_config will be set to the name of the image containing the change after execution, when the conversation finishes.
You can save that image name. For a new conversation, you can set "use_docker" to the saved name of the image to start execution there.
Agreed that this should be added to FAQ as I've answered this question a number of times. Any volunteers?
Originally posted by @sonichi in #877 (comment)
The "use_docker" arg in an agent's code_execution_config will be set to the name of the image containing the change after execution, when the conversation finishes.
You can save that image name. For a new conversation, you can set "use_docker" to the saved name of the image to start execution there.
Agreed that this should be added to FAQ as I've answered this question a number of times. Any volunteers?
Originally posted by @sonichi in #877 (comment)