--- language: - en license: apache-2.0 tags: - text-generation-inference - transformers - unsloth - trl - sft base_model: unsloth/Phi-3-medium-4k-instruct-bnb-4bit --- # TofuTune Phi-3 Medium TofuTune is designed to take a lightweight agile ticket describing a cloud infrastructure task and output valid terraform/openTofu code which completes the ticket. ### Limitations - Small initial dataset to POC the usecase. - Only trained for use with the Google Cloud provider. ### Format Prompt: ``` ## Ticket: Active Directory Domain Setup in GCP ## User Story: As a DevOps engineer, I want to deploy an Active Directory domain in GCP to centralize domain management for enhanced security and efficiency. ## Details: The task involves configuring a Google Managed Microsoft Active Directory domain. This setup will use a specific domain name and operate within the 'us-central1' region, with a reserved IP range to ensure secure and isolated network operations. ## Acceptance CriteriaDomain Configuration: - A Google Managed Active Directory domain named tfgen-{suffix}.org.com should be successfully created. - Location and IP Reservation: The domain should be located in us-central1 and must utilize the reserved IP range 192.168.255.0/24 for its operations. ``` Response: ``` resource "google_active_directory_domain" "ad-domain" { domain_name = "tfgen-${local.name_suffix}.org.com" locations = ["us-central1"] reserved_ip_range = "192.168.255.0/24" } ``` - **Developed by:** simonbutt - **License:** apache-2.0 - **Finetuned from model :** unsloth/Phi-3-medium-4k-instruct-bnb-4bit This phi-3 medium model was trained 2x faster with [Unsloth](https://github.com/unslothai/unsloth) and Huggingface's TRL library. [](https://github.com/unslothai/unsloth)