dockerize_ollama_models.yml

name: 'Dockerize Ollama models'

on:
  workflow_dispatch:
    inputs:
      ollama_version:
        description: 'Ollama version'
        required: true
        default: 'latest'

jobs:
  build:
    runs-on: ubuntu-latest
    strategy:
      matrix:
        models: [
          model: {
            name: 'all-minilm',
            tag: '22m',
            description: 'Embedding models on very large sentence level datasets'
          },
          model: {
            name: 'bge-m3',
            tag: '567m',
            description: 'BGE-M3 is a new model from BAAI distinguished for its versatility in Multi-Functionality, Multi-Linguality, and Multi-Granularity'
          },
          model: {
            name: 'codegemma',
            tag: '7b',
            description: 'CodeGemma is a collection of powerful, lightweight models that can perform a variety of coding tasks like fill-in-the-middle code completion, code generation, natural language understanding, mathematical reasoning, and instruction following'
          },
          model: {
            name: 'deepseek-coder',
            tag: '1.3b',
            description: 'DeepSeek Coder is a capable coding model trained on two trillion code and natural language tokens'
          },
          model: {
            name: 'gemma2',
            tag: '2b',
            description: 'Google Gemma 2 is a high-performing and efficient model available in three sizes: 2B, 9B, and 27B'
          },
          model: {
            name: 'llama3.1',
            tag: '8b',
            description: 'Llama 3.1 is a new state-of-the-art model from Meta available in 8B, 70B and 405B parameter sizes'
          },
          model: {
            name: 'llama3.2',
            tag: '1b',
            description: 'Llama 3.2 of Meta goes small with 1B and 3B models.'
          },
          model: {
            name: 'llama3.2',
            tag: '3b',
            description: 'Llama 3.2 of Meta goes small with 1B and 3B models.'
          },
          model: {
            name: 'llava',
            tag: '7b',
            description: 'LLaVA is a novel end-to-end trained large multimodal model that combines a vision encoder and Vicuna for general-purpose visual and language understanding. Updated to version 1.6'
          },
          model: {
            name: 'llava-phi3',
            tag: '3.8b',
            description: 'A new small LLaVA model fine-tuned from Phi 3 Mini'
          },
          model: {
            name: 'mistral',
            tag: '7b',
            description: 'The 7B model released by Mistral AI, updated to version 0.3'
          },
          model: {
            name: 'moondream',
            tag: '1.8b',
            description: 'moondream2 is a small vision language model designed to run efficiently on edge devices'
          },
          model: {
            name: 'mxbai-embed-large',
            tag: '335m',
            description: 'State-of-the-art large embedding model from mixedbread.ai'
          },
          model: {
            name: 'nomic-embed-text',
            tag: 'v1.5',
            description: 'A high-performing open embedding model with a large token context window'
          },
          model: {
            name: 'phi3',
            tag: '3.8b',
            description: 'Phi-3 is a family of lightweight 3B (Mini) and 14B (Medium) state-of-the-art open models by Microsoft'
          },
          model: {
            name: 'phi3.5',
            tag: '3.8b',
            description: 'A lightweight AI model with 3.8 billion parameters with performance overtaking similarly and larger sized models'
          },
          model: {
            name: 'qwen2',
            tag: '0.5b',
            description: 'Qwen2 is a new series of large language models from Alibaba group'
          },
          model: {
            name: 'qwen2.5',
            tag: '0.5b-instruct',
            description: 'Qwen2.5 models are pretrained on Alibaba latest large-scale dataset, encompassing up to 18 trillion tokens. The model supports up to 128K tokens and has multilingual support'
          },
          model: {
            name: 'qwen2.5-coder',
            tag: '1.5b',
            description: 'The latest series of Code-Specific Qwen models, with significant improvements in code generation, code reasoning, and code fixing.'
          },
          model: {
            name: 'snowflake-arctic-embed',
            tag: '335m',
            description: 'A suite of text embedding models by Snowflake, optimized for performance'
          },
          model: {
            name: 'starcoder2',
            tag: '3b',
            description: 'StarCoder2 is the next generation of transparently trained open code LLMs that comes in three sizes: 3B, 7B and 15B parameters'
          },
          model: {
            name: 'reader-lm',
            tag: '1.5b',
            description: 'A series of models that convert HTML content to Markdown content, which is useful for content conversion tasks'
          },
          model: {
            name: 'bespoke-minicheck',
            tag: '7b',
            description: 'A SOTA fact-checking model developed by Bespoke Labs'
          }
        ]
    steps:
      - name: Free Disk Space
        uses: jlumbroso/free-disk-space@54081f138730dfa15788a46383842cd2f914a1be # v1.3.1
        with:
          tool-cache: true

      - name: Checkout repository
        uses: actions/checkout@v4

      - name: Build and Deploy Docker Image
        uses: ./
        with:
          docker_username: ${{ secrets.DOCKER_USERNAME }}
          docker_password: ${{ secrets.DOCKER_PASSWORD }}
          docker_repository: ${{ matrix.models.model.name }}
          docker_tag: ${{ github.event.inputs.ollama_version }}-${{ matrix.models.model.tag }}
          model_name: ${{ matrix.models.model.name }}
          model_tag: ${{ matrix.models.model.tag }}
          ollama_version: ${{ github.event.inputs.ollama_version }}
          short_description: ${{ matrix.models.model.description }}