What are gpu requirements to run llama 2 13b on spaces

Though i use other cloud platform like aws and azure for deployment of apps ,here i found about hugging face spaces.I want to deploy my llama 13b customized model to hugging face spaces. what will be requirements or instances will be required for the deployment of 13b llama 2 to live with gradio.
also please clarify on the part where in future if ill get some issues with the server ,do we have support team for the help.i mean about the downtown ,do we get the downtime because i want to showcase this to the clients.