workflow_finetune.yml

name: Finetune

on:
  workflow_call:
    inputs:
      ci_type:
        type: string
        default: 'pr'
      runner_container_image:
        type: string
        default: '10.1.2.13:5000/llmray-build'
      http_proxy:
        type: string
        default: 'http://proxy-prc.intel.com:912'
      https_proxy:
        type: string
        default: 'http://proxy-prc.intel.com:912'
      runner_config_path:
        type: string
        default: '/home/ci/llm-ray-actions-runner'
      code_checkout_path:
        type: string
        default: '/home/ci/actions-runner/_work/llm-on-ray/llm-on-ray'
      model_cache_path:
        type: string
        default: '/mnt/DP_disk1/huggingface/cache'

concurrency:
  group: ${{ github.workflow }}-${{ github.event.pull_request.number || github.ref }}-ft
  cancel-in-progress: true

permissions:  # added using https://github.com/step-security/secure-repo
  contents: read

jobs:
  finetune:
    name: finetune
    strategy:
      matrix:
        model: [ EleutherAI/gpt-j-6b, NousResearch/Llama-2-7b-chat-hf, gpt2, bigscience/bloom-560m, facebook/opt-125m, mosaicml/mpt-7b, NousResearch/Llama-2-7b-hf, mistralai/Mistral-7B-v0.1]
        isPR:
          - ${{inputs.ci_type == 'pr'}}

        exclude:
          - { isPR: true }
        include:
          - { model: "EleutherAI/gpt-j-6b"}
          - { model: "NousResearch/Llama-2-7b-chat-hf"}
          - { model: "mistralai/Mistral-7B-v0.1"}

    runs-on: self-hosted

    defaults:
      run:
        shell: bash
    container:
      image: ${{ inputs.runner_container_image }}
      env:
        http_proxy: ${{ inputs.http_proxy }}
        https_proxy: ${{ inputs.https_proxy }}
        SHELL: bash -eo pipefail
      volumes:
        - /var/run/docker.sock:/var/run/docker.sock
        - ${{ inputs.runner_config_path }}:/root/actions-runner-config

    steps:
      - name: Checkout
        uses: actions/checkout@692973e3d937129bcbf40652eb9f2f61becf3332 # v4.1.7

      - name: Build Docker Image
        run: |
          DF_SUFFIX=".cpu_and_deepspeed"
          TARGET="finetune"
          USE_PROXY="1"
          source dev/scripts/ci-functions.sh
          build_and_prune ${TARGET} ${DF_SUFFIX} ${USE_PROXY}

      - name: Start Docker Container
        run: |
          TARGET="finetune"
          code_checkout_path=${{ inputs.code_checkout_path }}
          model_cache_path=${{ inputs.model_cache_path }}
          USE_PROXY="1"
          source dev/scripts/ci-functions.sh
          start_docker ${TARGET} ${code_checkout_path} ${model_cache_path} ${USE_PROXY}

      - name: Run Finetune Test
        run: |
          TARGET="finetune"
          source dev/scripts/ci-functions.sh
          finetune_test ${{ matrix.model }}

      - name: Run Finetune DPO Test
        run: |
          TARGET="finetune"
          source dev/scripts/ci-functions.sh
          finetune_dpo_test ${{ matrix.model }}

      - name: Run PEFT-LoRA Test
        run: |
          source dev/scripts/ci-functions.sh
          peft_lora_test ${{ matrix.model }}

      - name: Stop Ray
        run: |
          TARGET="finetune"
          source dev/scripts/ci-functions.sh
          stop_ray ${TARGET}

      - name: Stop Container
        if: success() || failure()
        run: |
          TARGET="finetune"
          source dev/scripts/ci-functions.sh
          stop_container ${TARGET}