The custom template container disk size is overwritten by python default code

[tleyden]

Describe the bug
When a custom template ID is provided, the python code will use the default value:

container_disk_in_gb:int=5

instead of using the value in the custom template.

A workaround is to do a graphql query to get the template containerInGb field, then pass that to runpod.create_pod().

    def get_custom_template(self, template_name: str):
        # Call the "me" api and ask for the list of templates
        QUERY_POD = """
            query myTemplates {
                myself {
                    podTemplates {
                        id
                        name
                        imageName
                        containerDiskInGb
                    }
                }
            }
            """
        result = run_graphql_query(QUERY_POD)
        pod_templates = result["data"]["myself"]["podTemplates"]
        for pod_template in pod_templates:
            if pod_template["name"] == template_name:
                return {
                    "containerDiskInGb": pod_template["containerDiskInGb"], 
                    "id:pod_template["id"], 
                    imageName:pod_template["imageName"]
                }
                return custom_template

        raise Exception("Could not find pod template id for template name: " + template_name)

Expected behavior
When a template ID is passed, use the container disk size specified in the template