change error handler behavior and fix error when a model isn't found

.dockerignore

@@ -1,9 +1,7 @@
+build
+llama/build
+.venv
 .vscode
 ollama
 app
-macapp
-dist
-llm/llama.cpp
-.env
-.cache
-test_data
+web

.gitattributes

@@ -1 +0,0 @@
-llm/ext_server/* linguist-vendored

.github/ISSUE_TEMPLATE/10_bug_report.yml

@@ -1,60 +0,0 @@
-name: Bug report
-labels: [bug]
-description: Something isn't working right.
-body:
-  - type: textarea
-    id: description
-    attributes:
-      label: What is the issue?
-      description: What happened? What did you expect to happen?
-    validations:
-      required: true
-  - type: dropdown
-    id: os
-    attributes:
-      label: OS
-      description: Which operating system are you using?
-      multiple: true
-      options:
-        - Linux
-        - macOS
-        - Windows
-        - Docker
-        - WSL2
-    validations:
-      required: false
-  - type: dropdown
-    id: gpu
-    attributes:
-      label: GPU
-      description: Which GPU are you using?
-      multiple: true
-      options:
-        - Nvidia
-        - AMD
-        - Intel
-        - Apple
-        - Other
-    validations:
-      required: false
-  - type: dropdown
-    id: cpu
-    attributes:
-      label: CPU
-      description: Which CPU are you using?
-      multiple: true
-      options:
-        - Intel
-        - AMD
-        - Apple
-        - Other
-    validations:
-      required: false
-  - type: input
-    id: version
-    attributes:
-      label: Ollama version
-      description: What version of Ollama are you using? (`ollama --version`)
-      placeholder: e.g., 0.1.32
-    validations:
-      required: false

.github/ISSUE_TEMPLATE/20_feature_request.md

@@ -1,6 +0,0 @@
----
-name: Feature request
-about: Request a new feature
-labels: feature request
----
-

.github/ISSUE_TEMPLATE/30_model_request.md

@@ -1,5 +0,0 @@
----
-name: Model request
-about: Request support for a new model to be added to Ollama
-labels: model request
----

.github/ISSUE_TEMPLATE/config.yml

@@ -1,8 +0,0 @@
-blank_issues_enabled: true
-contact_links:
-  - name: Help
-    url: https://discord.com/invite/ollama
-    about: Please join our Discord server for help using Ollama
-  - name: Troubleshooting
-    url: https://github.com/ollama/ollama/blob/main/docs/faq.md#faq
-    about: See the FAQ for common issues and solutions

.github/workflows/latest.yaml

@@ -1,24 +0,0 @@
-name: latest
-
-on:
-  release:
-    types: [released]
-
-jobs:
-  update-latest:
-    environment: release
-    runs-on: linux
-    steps:
-      - uses: actions/checkout@v4
-      - name: Login to Docker Hub
-        uses: docker/login-action@v3
-        with:
-          username: ${{ vars.DOCKER_USER }}
-          password: ${{ secrets.DOCKER_ACCESS_TOKEN }}
-      - name: Tag images as latest
-        env:
-          PUSH: "1"
-        shell: bash
-        run: |
-          export "VERSION=${GITHUB_REF_NAME#v}"
-          ./scripts/tag_latest.sh

.github/workflows/release.yaml

@@ -1,480 +0,0 @@
-name: release
-
-on:
-  push:
-    tags:
-      - 'v*'
-
-jobs:
-  # Full build of the Mac assets
-  build-darwin:
-    runs-on: macos-12
-    environment: release
-    steps:
-      - uses: actions/checkout@v4
-      - name: Set Version
-        shell: bash
-        run: |
-          echo "VERSION=${GITHUB_REF_NAME#v}" >> $GITHUB_ENV
-          echo "RELEASE_VERSION=$(echo ${GITHUB_REF_NAME} | cut -f1 -d-)" >> $GITHUB_ENV
-      - name: key
-        env:
-          MACOS_SIGNING_KEY: ${{ secrets.MACOS_SIGNING_KEY }}
-          MACOS_SIGNING_KEY_PASSWORD: ${{ secrets.MACOS_SIGNING_KEY_PASSWORD }}
-        run: |
-          echo $MACOS_SIGNING_KEY | base64 --decode > certificate.p12
-          security create-keychain -p password build.keychain
-          security default-keychain -s build.keychain
-          security unlock-keychain -p password build.keychain
-          security import certificate.p12 -k build.keychain -P $MACOS_SIGNING_KEY_PASSWORD -T /usr/bin/codesign
-          security set-key-partition-list -S apple-tool:,apple:,codesign: -s -k password build.keychain
-          security set-keychain-settings -lut 3600 build.keychain
-      - uses: actions/setup-go@v5
-        with:
-          go-version: "stable"
-          cache: true
-      - name: Build Darwin
-        env:
-          APPLE_IDENTITY: ${{ secrets.APPLE_IDENTITY }}
-          APPLE_PASSWORD: ${{ secrets.APPLE_PASSWORD }}
-          APPLE_TEAM_ID: ${{ vars.APPLE_TEAM_ID }}
-          APPLE_ID: ${{ vars.APPLE_ID }}
-          SDKROOT: /Applications/Xcode_13.4.1.app/Contents/Developer/Platforms/MacOSX.platform/Developer/SDKs/MacOSX.sdk
-          DEVELOPER_DIR: /Applications/Xcode_13.4.1.app/Contents/Developer
-        run: |
-          ./scripts/build_darwin.sh
-
-      - uses: actions/upload-artifact@v4
-        with:
-          name: dist-darwin
-          path: |
-            dist/*arwin*
-            !dist/*-cov
-
-  # Windows builds take a long time to both install the dependencies and build, so parallelize
-  # CPU generation step
-  generate-windows-cpu:
-    environment: release
-    runs-on: windows
-    env:
-      KEY_CONTAINER: ${{ vars.KEY_CONTAINER }}
-    steps:
-      - uses: actions/checkout@v4
-      - name: Set Version
-        shell: bash
-        run: echo "VERSION=${GITHUB_REF_NAME#v}" >> $GITHUB_ENV
-      - uses: 'google-github-actions/auth@v2'
-        with:
-          project_id: 'ollama'
-          credentials_json: '${{ secrets.GOOGLE_SIGNING_CREDENTIALS }}'
-      - run: echo "${{ vars.OLLAMA_CERT }}" > ollama_inc.crt
-      - name: install Windows SDK 8.1 to get signtool
-        run: |
-          $ErrorActionPreference = "Stop"
-          write-host "downloading SDK"
-          Invoke-WebRequest -Uri "https://go.microsoft.com/fwlink/p/?LinkId=323507" -OutFile "${env:RUNNER_TEMP}\sdksetup.exe"
-          Start-Process "${env:RUNNER_TEMP}\sdksetup.exe" -ArgumentList @("/q") -NoNewWindow -Wait
-          write-host "Win SDK 8.1 installed"
-          gci -path 'C:\Program Files (x86)\Windows Kits\' -r -fi 'signtool.exe'
-      - name: install signing plugin
-        run: |
-          $ErrorActionPreference = "Stop"
-          write-host "downloading plugin"
-          Invoke-WebRequest -Uri "https://github.com/GoogleCloudPlatform/kms-integrations/releases/download/cng-v1.0/kmscng-1.0-windows-amd64.zip" -OutFile "${env:RUNNER_TEMP}\plugin.zip"
-          Expand-Archive -Path "${env:RUNNER_TEMP}\plugin.zip" -DestinationPath ${env:RUNNER_TEMP}\plugin\
-          write-host "Installing plugin"
-          & "${env:RUNNER_TEMP}\plugin\*\kmscng.msi" /quiet
-          write-host "plugin installed"
-      - uses: actions/setup-go@v5
-        with:
-          go-version: "stable"
-          cache: true
-      - run: go get ./...
-      - run: |
-          $gopath=(get-command go).source | split-path -parent
-          & "C:\Program Files (x86)\Microsoft Visual Studio\2019\Enterprise\Common7\Tools\Launch-VsDevShell.ps1"
-          cd $env:GITHUB_WORKSPACE
-          $env:CMAKE_SYSTEM_VERSION="10.0.22621.0"
-          $env:PATH="$gopath;$env:PATH"
-          go generate -x ./...
-        name: go generate
-      - uses: actions/upload-artifact@v4
-        with:
-          name: generate-windows-cpu
-          path: |
-            llm/build/**/bin/*
-            llm/build/**/*.a
-            dist/windows-amd64/**
-
-  # ROCm generation step
-  generate-windows-rocm:
-    environment: release
-    runs-on: windows
-    env:
-      KEY_CONTAINER: ${{ vars.KEY_CONTAINER }}
-    steps:
-      - uses: actions/checkout@v4
-      - name: Set Version
-        shell: bash
-        run: echo "VERSION=${GITHUB_REF_NAME#v}" >> $GITHUB_ENV
-      - uses: 'google-github-actions/auth@v2'
-        with:
-          project_id: 'ollama'
-          credentials_json: '${{ secrets.GOOGLE_SIGNING_CREDENTIALS }}'
-      - run: echo "${{ vars.OLLAMA_CERT }}" > ollama_inc.crt
-      - name: install Windows SDK 8.1 to get signtool
-        run: |
-          $ErrorActionPreference = "Stop"
-          write-host "downloading SDK"
-          Invoke-WebRequest -Uri "https://go.microsoft.com/fwlink/p/?LinkId=323507" -OutFile "${env:RUNNER_TEMP}\sdksetup.exe"
-          Start-Process "${env:RUNNER_TEMP}\sdksetup.exe" -ArgumentList @("/q") -NoNewWindow -Wait
-          write-host "Win SDK 8.1 installed"
-          gci -path 'C:\Program Files (x86)\Windows Kits\' -r -fi 'signtool.exe'
-      - name: install signing plugin
-        run: |
-          $ErrorActionPreference = "Stop"
-          write-host "downloading plugin"
-          Invoke-WebRequest -Uri "https://github.com/GoogleCloudPlatform/kms-integrations/releases/download/cng-v1.0/kmscng-1.0-windows-amd64.zip" -OutFile "${env:RUNNER_TEMP}\plugin.zip"
-          Expand-Archive -Path "${env:RUNNER_TEMP}\plugin.zip" -DestinationPath ${env:RUNNER_TEMP}\plugin\
-          write-host "Installing plugin"
-          & "${env:RUNNER_TEMP}\plugin\*\kmscng.msi" /quiet
-          write-host "plugin installed"
-      - uses: actions/setup-go@v5
-        with:
-          go-version: "stable"
-          cache: true
-      - name: 'Install ROCm'
-        run: |
-          $ErrorActionPreference = "Stop"
-          write-host "downloading AMD HIP Installer"
-          Invoke-WebRequest -Uri "https://download.amd.com/developer/eula/rocm-hub/AMD-Software-PRO-Edition-24.Q3-WinSvr2022-For-HIP.exe" -OutFile "${env:RUNNER_TEMP}\rocm-install.exe"
-          write-host "Installing AMD HIP"
-          Start-Process "${env:RUNNER_TEMP}\rocm-install.exe" -ArgumentList '-install' -NoNewWindow -Wait
-          write-host "Completed AMD HIP"
-      - name: 'Verify ROCm'
-        run: |
-          & 'C:\Program Files\AMD\ROCm\*\bin\clang.exe' --version
-      - run: go get ./...
-      - run: |
-          $gopath=(get-command go).source | split-path -parent
-          & "C:\Program Files (x86)\Microsoft Visual Studio\2019\Enterprise\Common7\Tools\Launch-VsDevShell.ps1"
-          cd $env:GITHUB_WORKSPACE
-          $env:CMAKE_SYSTEM_VERSION="10.0.22621.0"
-          $env:PATH="$gopath;$env:PATH"
-          $env:OLLAMA_SKIP_CPU_GENERATE="1"
-          $env:HIP_PATH=$(Resolve-Path 'C:\Program Files\AMD\ROCm\*\bin\clang.exe' | split-path | split-path)
-          go generate -x ./...
-        name: go generate
-      - name: 'gather rocm dependencies'
-        run: |
-          $HIP_PATH=$(Resolve-Path 'C:\Program Files\AMD\ROCm\*\bin\clang.exe' | split-path | split-path)
-          md "dist\deps\bin\rocblas\library"
-          cp "${HIP_PATH}\bin\hipblas.dll" "dist\deps\bin\"
-          cp "${HIP_PATH}\bin\rocblas.dll" "dist\deps\bin\"
-          cp "${HIP_PATH}\bin\rocblas\library\*" "dist\deps\bin\rocblas\library\"
-      - uses: actions/upload-artifact@v4
-        with:
-          name: generate-windows-rocm
-          path: |
-            llm/build/**/bin/*
-            dist/windows-amd64/**
-      - uses: actions/upload-artifact@v4
-        with:
-          name: windows-rocm-deps
-          path: dist/deps/*
-
-  # CUDA generation step
-  generate-windows-cuda:
-    environment: release
-    runs-on: windows
-    env:
-      KEY_CONTAINER: ${{ vars.KEY_CONTAINER }}
-    steps:
-      - uses: actions/checkout@v4
-      - name: Set Version
-        shell: bash
-        run: echo "VERSION=${GITHUB_REF_NAME#v}" >> $GITHUB_ENV
-      - uses: 'google-github-actions/auth@v2'
-        with:
-          project_id: 'ollama'
-          credentials_json: '${{ secrets.GOOGLE_SIGNING_CREDENTIALS }}'
-      - run: echo "${{ vars.OLLAMA_CERT }}" > ollama_inc.crt
-      - name: install Windows SDK 8.1 to get signtool
-        run: |
-          $ErrorActionPreference = "Stop"
-          write-host "downloading SDK"
-          Invoke-WebRequest -Uri "https://go.microsoft.com/fwlink/p/?LinkId=323507" -OutFile "${env:RUNNER_TEMP}\sdksetup.exe"
-          Start-Process "${env:RUNNER_TEMP}\sdksetup.exe" -ArgumentList @("/q") -NoNewWindow -Wait
-          write-host "Win SDK 8.1 installed"
-          gci -path 'C:\Program Files (x86)\Windows Kits\' -r -fi 'signtool.exe'
-      - name: install signing plugin
-        run: |
-          $ErrorActionPreference = "Stop"
-          write-host "downloading plugin"
-          Invoke-WebRequest -Uri "https://github.com/GoogleCloudPlatform/kms-integrations/releases/download/cng-v1.0/kmscng-1.0-windows-amd64.zip" -OutFile "${env:RUNNER_TEMP}\plugin.zip"
-          Expand-Archive -Path "${env:RUNNER_TEMP}\plugin.zip" -DestinationPath ${env:RUNNER_TEMP}\plugin\
-          write-host "Installing plugin"
-          & "${env:RUNNER_TEMP}\plugin\*\kmscng.msi" /quiet
-          write-host "plugin installed"
-      - uses: actions/setup-go@v5
-        with:
-          go-version: "stable"
-          cache: true
-      - name: 'Install CUDA'
-        run: |
-          $ErrorActionPreference = "Stop"
-          write-host "downloading CUDA Installer"
-          Invoke-WebRequest -Uri "https://developer.download.nvidia.com/compute/cuda/11.3.1/local_installers/cuda_11.3.1_465.89_win10.exe" -OutFile "${env:RUNNER_TEMP}\cuda-install.exe"
-          write-host "Installing CUDA"
-          Start-Process "${env:RUNNER_TEMP}\cuda-install.exe" -ArgumentList '-s' -NoNewWindow -Wait
-          write-host "Completed CUDA"
-          $cudaPath=((resolve-path "c:\Program Files\NVIDIA*\CUDA\v*\bin\nvcc.exe")[0].path | split-path | split-path)
-          $cudaVer=($cudaPath | split-path -leaf ) -replace 'v(\d+).(\d+)', '$1_$2' 
-          echo "$cudaPath\bin" >> $env:GITHUB_PATH
-          echo "CUDA_PATH=$cudaPath" >> $env:GITHUB_ENV
-          echo "CUDA_PATH_V${cudaVer}=$cudaPath" >> $env:GITHUB_ENV
-          echo "CUDA_PATH_VX_Y=CUDA_PATH_V${cudaVer}" >> $env:GITHUB_ENV
-      - name: 'Verify CUDA'
-        run: nvcc -V
-      - run: go get ./...
-      - name: go generate
-        run: |
-          $gopath=(get-command go).source | split-path -parent
-          $cudabin=(get-command nvcc).source | split-path
-          & "C:\Program Files (x86)\Microsoft Visual Studio\2019\Enterprise\Common7\Tools\Launch-VsDevShell.ps1"
-          cd $env:GITHUB_WORKSPACE
-          $env:CMAKE_SYSTEM_VERSION="10.0.22621.0"
-          $env:PATH="$gopath;$cudabin;$env:PATH"
-          $env:OLLAMA_SKIP_CPU_GENERATE="1"
-          go generate -x ./...
-      - name: 'gather cuda dependencies'
-        run: |
-          $NVIDIA_DIR=(resolve-path 'C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\*\bin\')[0]
-          md "dist\deps"
-          cp "${NVIDIA_DIR}\cudart64_*.dll" "dist\deps\"
-          cp "${NVIDIA_DIR}\cublas64_*.dll" "dist\deps\"
-          cp "${NVIDIA_DIR}\cublasLt64_*.dll" "dist\deps\"
-      - uses: actions/upload-artifact@v4
-        with:
-          name: generate-windows-cuda
-          path: |
-            llm/build/**/bin/*
-            dist/windows-amd64/**
-      - uses: actions/upload-artifact@v4
-        with:
-          name: windows-cuda-deps
-          path: dist/deps/*
-
-  # Import the prior generation steps and build the final windows assets
-  build-windows:
-    environment: release
-    runs-on: windows
-    needs:
-      - generate-windows-cuda
-      - generate-windows-rocm
-      - generate-windows-cpu
-    env:
-      KEY_CONTAINER: ${{ vars.KEY_CONTAINER }}
-    steps:
-      - uses: actions/checkout@v4
-        with:
-          submodules: recursive
-      - name: Set Version
-        shell: bash
-        run: echo "VERSION=${GITHUB_REF_NAME#v}" >> $GITHUB_ENV
-      - uses: 'google-github-actions/auth@v2'
-        with:
-          project_id: 'ollama'
-          credentials_json: '${{ secrets.GOOGLE_SIGNING_CREDENTIALS }}'
-      - run: echo "${{ vars.OLLAMA_CERT }}" > ollama_inc.crt
-      - name: install Windows SDK 8.1 to get signtool
-        run: |
-          $ErrorActionPreference = "Stop"
-          write-host "downloading SDK"
-          Invoke-WebRequest -Uri "https://go.microsoft.com/fwlink/p/?LinkId=323507" -OutFile "${env:RUNNER_TEMP}\sdksetup.exe"
-          Start-Process "${env:RUNNER_TEMP}\sdksetup.exe" -ArgumentList @("/q") -NoNewWindow -Wait
-          write-host "Win SDK 8.1 installed"
-          gci -path 'C:\Program Files (x86)\Windows Kits\' -r -fi 'signtool.exe'
-      - name: install signing plugin
-        run: |
-          $ErrorActionPreference = "Stop"
-          write-host "downloading plugin"
-          Invoke-WebRequest -Uri "https://github.com/GoogleCloudPlatform/kms-integrations/releases/download/cng-v1.0/kmscng-1.0-windows-amd64.zip" -OutFile "${env:RUNNER_TEMP}\plugin.zip"
-          Expand-Archive -Path "${env:RUNNER_TEMP}\plugin.zip" -DestinationPath ${env:RUNNER_TEMP}\plugin\
-          write-host "Installing plugin"
-          & "${env:RUNNER_TEMP}\plugin\*\kmscng.msi" /quiet
-          write-host "plugin installed"
-      - uses: actions/setup-go@v5
-        with:
-          go-version: "stable"
-          cache: true
-      - run: go get
-      - uses: actions/download-artifact@v4
-        with:
-          name: generate-windows-cpu
-      - uses: actions/download-artifact@v4
-        with:
-          name: generate-windows-cuda
-      - uses: actions/download-artifact@v4
-        with:
-          name: windows-cuda-deps
-      - uses: actions/download-artifact@v4
-        with:
-          name: windows-rocm-deps
-      - uses: actions/download-artifact@v4
-        with:
-          name: generate-windows-rocm
-      - run: dir llm/build
-      - run: |
-          $gopath=(get-command go).source | split-path -parent
-          & "C:\Program Files (x86)\Microsoft Visual Studio\2019\Enterprise\Common7\Tools\Launch-VsDevShell.ps1"
-          cd $env:GITHUB_WORKSPACE
-          $env:CMAKE_SYSTEM_VERSION="10.0.22621.0"
-          $env:PATH="$gopath;$env:PATH"
-          $env:OLLAMA_SKIP_GENERATE="1"
-          & .\scripts\build_windows.ps1
-      - uses: actions/upload-artifact@v4
-        with:
-          name: dist-windows
-          path: |
-            dist/OllamaSetup.exe
-            dist/ollama-windows-*.zip
-
-  # Linux x86 assets built using the container based build
-  build-linux-amd64:
-    environment: release
-    runs-on: linux
-    env:
-      OLLAMA_SKIP_MANIFEST_CREATE: '1'
-      BUILD_ARCH: amd64
-      PUSH: '1'
-    steps:
-      - uses: actions/checkout@v4
-        with:
-          submodules: recursive
-      - name: Set Version
-        shell: bash
-        run: echo "VERSION=${GITHUB_REF_NAME#v}" >> $GITHUB_ENV
-      - name: Login to Docker Hub
-        uses: docker/login-action@v3
-        with:
-          username: ${{ vars.DOCKER_USER }}
-          password: ${{ secrets.DOCKER_ACCESS_TOKEN }}
-      - run: |
-          ./scripts/build_linux.sh
-          ./scripts/build_docker.sh
-          mv dist/deps/* dist/
-      - uses: actions/upload-artifact@v4
-        with:
-          name: dist-linux-amd64
-          path: |
-            dist/*linux*
-            !dist/*-cov
-
-  # Linux ARM assets built using the container based build
-  # (at present, docker isn't pre-installed on arm ubunutu images)
-  build-linux-arm64:
-    environment: release
-    runs-on: linux-arm64
-    env:
-      OLLAMA_SKIP_MANIFEST_CREATE: '1'
-      BUILD_ARCH: arm64
-      PUSH: '1'
-    steps:
-      - uses: actions/checkout@v4
-        with:
-          submodules: recursive
-      - name: Set Version
-        shell: bash
-        run: echo "VERSION=${GITHUB_REF_NAME#v}" >> $GITHUB_ENV
-      - name: 'Install Docker'
-        run: |
-          # Add Docker's official GPG key:
-          env
-          uname -a
-          sudo apt-get update
-          sudo apt-get install -y ca-certificates curl
-          sudo install -m 0755 -d /etc/apt/keyrings
-          sudo curl -fsSL https://download.docker.com/linux/ubuntu/gpg -o /etc/apt/keyrings/docker.asc
-          sudo chmod a+r /etc/apt/keyrings/docker.asc
-
-          # Add the repository to Apt sources:
-          echo \
-            "deb [arch=$(dpkg --print-architecture) signed-by=/etc/apt/keyrings/docker.asc] https://download.docker.com/linux/ubuntu \
-            $(. /etc/os-release && echo "$VERSION_CODENAME") stable" | \
-            sudo tee /etc/apt/sources.list.d/docker.list > /dev/null
-          sudo apt-get update
-          sudo apt-get install -y docker-ce docker-ce-cli containerd.io
-          sudo usermod -aG docker $USER
-          sudo apt-get install acl
-          sudo setfacl --modify user:$USER:rw /var/run/docker.sock
-      - name: Login to Docker Hub
-        uses: docker/login-action@v3
-        with:
-          username: ${{ vars.DOCKER_USER }}
-          password: ${{ secrets.DOCKER_ACCESS_TOKEN }}
-      - run: |
-          ./scripts/build_linux.sh
-          ./scripts/build_docker.sh
-      - uses: actions/upload-artifact@v4
-        with:
-          name: dist-linux-arm64
-          path: |
-            dist/*linux*
-            !dist/*-cov
-
-  # Aggregate all the assets and ship a release
-  release:
-    needs:
-      - build-darwin
-      - build-windows
-      - build-linux-amd64
-      - build-linux-arm64
-    runs-on: linux
-    environment: release
-    permissions:
-      contents: write
-    env:
-      OLLAMA_SKIP_IMAGE_BUILD: '1'
-      PUSH: '1'
-      GH_TOKEN: ${{ github.token }}
-    steps:
-      - uses: actions/checkout@v4
-      - name: Set Version
-        shell: bash
-        run: |
-          echo "VERSION=${GITHUB_REF_NAME#v}" >> $GITHUB_ENV
-          echo "RELEASE_VERSION=$(echo ${GITHUB_REF_NAME} | cut -f1 -d-)" >> $GITHUB_ENV
-      - name: Login to Docker Hub
-        uses: docker/login-action@v3
-        with:
-          username: ${{ vars.DOCKER_USER }}
-          password: ${{ secrets.DOCKER_ACCESS_TOKEN }}
-      - run: ./scripts/build_docker.sh
-      - name: Retrieve built artifact
-        uses: actions/download-artifact@v4
-        with:
-          path: dist
-          pattern: dist-*
-          merge-multiple: true
-      - run: |
-          ls -lh dist/
-          (cd dist; sha256sum * > sha256sum.txt)
-          cat dist/sha256sum.txt
-      - name: Create or update Release
-        run: |
-          echo "Looking for existing release for ${{ env.RELEASE_VERSION }}"
-          OLD_TAG=$(gh release ls --json name,tagName | jq -r ".[] | select(.name == \"${{ env.RELEASE_VERSION }}\") | .tagName")
-          if [ -n "$OLD_TAG" ]; then
-            echo "Updating release ${{ env.RELEASE_VERSION }} to point to new tag ${GITHUB_REF_NAME}"
-            gh release edit ${OLD_TAG} --tag ${GITHUB_REF_NAME}
-          else
-            echo "Creating new release ${{ env.RELEASE_VERSION }} pointing to tag ${GITHUB_REF_NAME}"
-            gh release create ${GITHUB_REF_NAME} \
-              --title ${{ env.RELEASE_VERSION }} \
-              --draft \
-              --generate-notes \
-              --prerelease
-          fi
-          echo "Uploading artifacts for tag ${GITHUB_REF_NAME}"
-          gh release upload ${GITHUB_REF_NAME} dist/* --clobber

.github/workflows/test.yaml

@@ -1,323 +0,0 @@
-name: test
-
-concurrency:
-  # For PRs, later CI runs preempt previous ones. e.g. a force push on a PR
-  # cancels running CI jobs and starts all new ones.
-  #
-  # For non-PR pushes, concurrency.group needs to be unique for every distinct
-  # CI run we want to have happen. Use run_id, which in practice means all
-  # non-PR CI runs will be allowed to run without preempting each other.
-  group: ${{ github.workflow }}-$${{ github.pull_request.number || github.run_id }}
-  cancel-in-progress: true
-
-on:
-  pull_request:
-    paths:
-      - '**/*'
-      - '!docs/**'
-      - '!README.md'
-
-jobs:
-  changes:
-    runs-on: ubuntu-latest
-    outputs:
-      GENERATE: ${{ steps.changes.outputs.GENERATE }}
-      GENERATE_CUDA: ${{ steps.changes.outputs.GENERATE_CUDA }}
-      GENERATE_ROCM: ${{ steps.changes.outputs.GENERATE_ROCM }}
-    steps:
-      - uses: actions/checkout@v4
-        with:
-          fetch-depth: 0
-      - id: changes
-        run: |
-          changed() {
-            git diff-tree -r --no-commit-id --name-only \
-              $(git merge-base ${{ github.event.pull_request.base.sha }} ${{ github.event.pull_request.head.sha }}) \
-              ${{ github.event.pull_request.head.sha }} \
-              | xargs python3 -c "import sys; from pathlib import Path; print(any(Path(x).match(glob) for x in sys.argv[1:] for glob in '$*'.split(' ')))"
-          }
-
-          {
-            echo GENERATE=$(changed 'llm/llama.cpp' 'llm/patches/**' 'llm/ext_server/**' 'llm/generate/**')
-            echo GENERATE_CUDA=$(changed 'llm/llama.cpp' 'llm/patches/**' 'llm/ext_server/**' 'llm/generate/**')
-            echo GENERATE_ROCM=$(changed 'llm/llama.cpp' 'llm/patches/**' 'llm/ext_server/**' 'llm/generate/**')
-          } >>$GITHUB_OUTPUT
-
-  generate:
-    needs: [changes]
-    if: ${{ needs.changes.outputs.GENERATE == 'True' }}
-    strategy:
-      matrix:
-        os: [ubuntu-latest, macos-latest, windows-2019]
-        arch: [amd64, arm64]
-        exclude:
-          - os: ubuntu-latest
-            arch: arm64
-          - os: windows-2019
-            arch: arm64
-    runs-on: ${{ matrix.os }}
-    env:
-      GOARCH: ${{ matrix.arch }}
-      CGO_ENABLED: '1'
-    steps:
-      - uses: actions/checkout@v4
-      - uses: actions/setup-go@v5
-        with:
-          go-version: "stable"
-          cache: true
-      - run: go get ./...
-      - run: |
-          $gopath=(get-command go).source | split-path -parent
-          $gccpath=(get-command gcc).source | split-path -parent
-          & "C:\Program Files (x86)\Microsoft Visual Studio\2019\Enterprise\Common7\Tools\Launch-VsDevShell.ps1"
-          cd $env:GITHUB_WORKSPACE
-          $env:CMAKE_SYSTEM_VERSION="10.0.22621.0"
-          $env:PATH="$gopath;$gccpath;$env:PATH"
-          echo $env:PATH
-          go generate -x ./...
-        if: ${{ startsWith(matrix.os, 'windows-') }}
-        name: 'Windows Go Generate'
-      - run: go generate -x ./...
-        if: ${{ ! startsWith(matrix.os, 'windows-') }}
-        name: 'Unix Go Generate'
-      - run: go build .
-      - uses: actions/upload-artifact@v4
-        with:
-          name: ${{ matrix.os }}-${{ matrix.arch }}-libraries
-          path: |
-            llm/build/**/bin/*
-            llm/build/**/*.a
-  generate-cuda:
-    needs: [changes]
-    if: ${{ needs.changes.outputs.GENERATE_CUDA == 'True' }}
-    strategy:
-      matrix:
-        cuda-version:
-          - '11.8.0'
-    runs-on: linux
-    container: nvidia/cuda:${{ matrix.cuda-version }}-devel-ubuntu20.04
-    steps:
-      - run: |
-          apt-get update && apt-get install -y git build-essential curl
-          curl -fsSL https://github.com/Kitware/CMake/releases/download/v3.28.1/cmake-3.28.1-linux-x86_64.tar.gz \
-            | tar -zx -C /usr --strip-components 1
-        env:
-          DEBIAN_FRONTEND: noninteractive
-      - uses: actions/checkout@v4
-      - uses: actions/setup-go@v4
-        with:
-          go-version-file: go.mod
-          cache: true
-      - run: go get ./...
-      - run: |
-          git config --global --add safe.directory /__w/ollama/ollama
-          go generate -x ./...
-        env:
-          OLLAMA_SKIP_CPU_GENERATE: '1'
-      - uses: actions/upload-artifact@v4
-        with:
-          name: cuda-${{ matrix.cuda-version }}-libraries
-          path: |
-            llm/build/**/bin/*
-            dist/windows-amd64/**
-  generate-rocm:
-    needs: [changes]
-    if: ${{ needs.changes.outputs.GENERATE_ROCM == 'True' }}
-    strategy:
-      matrix:
-        rocm-version:
-          - '6.1.2'
-    runs-on: linux
-    container: rocm/dev-ubuntu-20.04:${{ matrix.rocm-version }}
-    steps:
-      - run: |
-          apt-get update && apt-get install -y git build-essential curl rocm-libs
-          curl -fsSL https://github.com/Kitware/CMake/releases/download/v3.28.1/cmake-3.28.1-linux-x86_64.tar.gz \
-            | tar -zx -C /usr --strip-components 1
-        env:
-          DEBIAN_FRONTEND: noninteractive
-      - uses: actions/checkout@v4
-      - uses: actions/setup-go@v4
-        with:
-          go-version-file: go.mod
-          cache: true
-      - run: go get ./...
-      - run: |
-          git config --global --add safe.directory /__w/ollama/ollama
-          go generate -x ./...
-        env:
-          OLLAMA_SKIP_CPU_GENERATE: '1'
-      - uses: actions/upload-artifact@v4
-        with:
-          name: rocm-${{ matrix.rocm-version }}-libraries
-          path: |
-            llm/build/**/bin/*
-            dist/windows-amd64/**
-
-  # ROCm generation step
-  generate-windows-rocm:
-    needs: [changes]
-    if: ${{ needs.changes.outputs.GENERATE_ROCM == 'True' }}
-    runs-on: windows
-    steps:
-      - uses: actions/checkout@v4
-      - uses: actions/setup-go@v5
-        with:
-          go-version: "stable"
-          cache: true
-      - name: 'Install ROCm'
-        run: |
-          $ErrorActionPreference = "Stop"
-          write-host "downloading AMD HIP Installer"
-          Invoke-WebRequest -Uri "https://download.amd.com/developer/eula/rocm-hub/AMD-Software-PRO-Edition-24.Q3-WinSvr2022-For-HIP.exe" -OutFile "${env:RUNNER_TEMP}\rocm-install.exe"
-          write-host "Installing AMD HIP"
-          Start-Process "${env:RUNNER_TEMP}\rocm-install.exe" -ArgumentList '-install' -NoNewWindow -Wait
-          write-host "Completed AMD HIP"
-      - name: 'Verify ROCm'
-        run: |
-          & 'C:\Program Files\AMD\ROCm\*\bin\clang.exe' --version
-      - run: go get ./...
-      - run: |
-          $gopath=(get-command go).source | split-path -parent
-          & "C:\Program Files (x86)\Microsoft Visual Studio\2019\Enterprise\Common7\Tools\Launch-VsDevShell.ps1"
-          cd $env:GITHUB_WORKSPACE
-          $env:CMAKE_SYSTEM_VERSION="10.0.22621.0"
-          $env:PATH="$gopath;$env:PATH"
-          $env:OLLAMA_SKIP_CPU_GENERATE="1"
-          $env:HIP_PATH=$(Resolve-Path 'C:\Program Files\AMD\ROCm\*\bin\clang.exe' | split-path | split-path)
-          go generate -x ./...
-        name: go generate
-        env:
-          OLLAMA_SKIP_CPU_GENERATE: '1'
-      # TODO - do we need any artifacts?
-
-  # CUDA generation step
-  generate-windows-cuda:
-    needs: [changes]
-    if: ${{ needs.changes.outputs.GENERATE_CUDA == 'True' }}
-    runs-on: windows
-    steps:
-      - uses: actions/checkout@v4
-      - uses: actions/setup-go@v5
-        with:
-          go-version: "stable"
-          cache: true
-      - name: 'Install CUDA'
-        run: |
-          $ErrorActionPreference = "Stop"
-          write-host "downloading CUDA Installer"
-          Invoke-WebRequest -Uri "https://developer.download.nvidia.com/compute/cuda/11.3.1/local_installers/cuda_11.3.1_465.89_win10.exe" -OutFile "${env:RUNNER_TEMP}\cuda-install.exe"
-          write-host "Installing CUDA"
-          Start-Process "${env:RUNNER_TEMP}\cuda-install.exe" -ArgumentList '-s' -NoNewWindow -Wait
-          write-host "Completed CUDA"
-          $cudaPath=((resolve-path "c:\Program Files\NVIDIA*\CUDA\v*\bin\nvcc.exe")[0].path | split-path | split-path)
-          $cudaVer=($cudaPath | split-path -leaf ) -replace 'v(\d+).(\d+)', '$1_$2' 
-          echo "$cudaPath\bin" >> $env:GITHUB_PATH
-          echo "CUDA_PATH=$cudaPath" >> $env:GITHUB_ENV
-          echo "CUDA_PATH_V${cudaVer}=$cudaPath" >> $env:GITHUB_ENV
-          echo "CUDA_PATH_VX_Y=CUDA_PATH_V${cudaVer}" >> $env:GITHUB_ENV
-      - name: 'Verify CUDA'
-        run: nvcc -V
-      - run: go get ./...
-      - name: go generate
-        run: |
-          $gopath=(get-command go).source | split-path -parent
-          $cudabin=(get-command nvcc).source | split-path
-          & "C:\Program Files (x86)\Microsoft Visual Studio\2019\Enterprise\Common7\Tools\Launch-VsDevShell.ps1"
-          cd $env:GITHUB_WORKSPACE
-          $env:CMAKE_SYSTEM_VERSION="10.0.22621.0"
-          $env:PATH="$gopath;$cudabin;$env:PATH"
-          $env:OLLAMA_SKIP_CPU_GENERATE="1"
-          go generate -x ./...
-        env:
-          OLLAMA_SKIP_CPU_GENERATE: '1'
-      # TODO - do we need any artifacts?
-
-  lint:
-    strategy:
-      matrix:
-        os: [ubuntu-latest, macos-latest, windows-2019]
-        arch: [amd64, arm64]
-        exclude:
-          - os: ubuntu-latest
-            arch: arm64
-          - os: windows-2019
-            arch: arm64
-          - os: macos-latest
-            arch: amd64
-    runs-on: ${{ matrix.os }}
-    env:
-      GOARCH: ${{ matrix.arch }}
-      CGO_ENABLED: '1'
-    steps:
-      - uses: actions/checkout@v4
-        with:
-          submodules: recursive
-      - uses: actions/setup-go@v5
-        with:
-          go-version: "stable"
-          cache: false
-      - run: |
-          case ${{ matrix.arch }} in
-            amd64) echo ARCH=x86_64 ;;
-            arm64) echo ARCH=arm64 ;;
-          esac >>$GITHUB_ENV
-        shell: bash
-      - run: |
-          mkdir -p llm/build/linux/$ARCH/stub/bin
-          touch llm/build/linux/$ARCH/stub/bin/ollama_llama_server
-        if: ${{ startsWith(matrix.os, 'ubuntu-') }}
-      - run: |
-          mkdir -p llm/build/darwin/$ARCH/stub/bin
-          touch llm/build/darwin/$ARCH/stub/bin/ollama_llama_server
-        if: ${{ startsWith(matrix.os, 'macos-') }}
-      - uses: golangci/golangci-lint-action@v6
-        with:
-          args: --timeout 8m0s -v ${{ startsWith(matrix.os, 'windows-') && '' || '--disable gofmt --disable goimports' }}
-  test:
-    strategy:
-      matrix:
-        os: [ubuntu-latest, macos-latest, windows-2019]
-        arch: [amd64]
-        exclude:
-          - os: ubuntu-latest
-            arch: arm64
-          - os: windows-2019
-            arch: arm64
-    runs-on: ${{ matrix.os }}
-    env:
-      GOARCH: ${{ matrix.arch }}
-      CGO_ENABLED: '1'
-      OLLAMA_CPU_TARGET: 'static'
-      OLLAMA_SKIP_CPU_GENERATE: '1'
-      OLLAMA_SKIP_METAL_GENERATE: '1'
-    steps:
-      - uses: actions/checkout@v4
-        with:
-          submodules: recursive
-      - uses: actions/setup-go@v5
-        with:
-          go-version: "stable"
-          cache: true
-      - run: |
-          case ${{ matrix.arch }} in
-            amd64) echo ARCH=x86_64 ;;
-            arm64) echo ARCH=arm64 ;;
-          esac >>$GITHUB_ENV
-        shell: bash
-      - run: |
-          mkdir -p llm/build/linux/$ARCH/stub/bin
-          touch llm/build/linux/$ARCH/stub/bin/ollama_llama_server
-        if: ${{ startsWith(matrix.os, 'ubuntu-') }}
-      - run: |
-          mkdir -p llm/build/darwin/$ARCH/stub/bin
-          touch llm/build/darwin/$ARCH/stub/bin/ollama_llama_server
-        if: ${{ startsWith(matrix.os, 'macos-') }}
-        shell: bash
-      - run: go generate ./...
-      - run: go build
-      - run: go test -v ./...
-      - uses: actions/upload-artifact@v4
-        with:
-          name: ${{ matrix.os }}-binaries
-          path: ollama

.gitignore

@@ -2,14 +2,5 @@
 .vscode
 .env
 .venv
-.swp
 dist
 ollama
-ggml-metal.metal
-.cache
-*.exe
-.idea
-test_data
-*.crt
-llm/build
-__debug_bin*

.gitmodules

@@ -1,4 +0,0 @@
-[submodule "llama.cpp"]
-	path = llm/llama.cpp
-	url = https://github.com/ggerganov/llama.cpp.git
-	shallow = true

.golangci.yaml

@@ -1,34 +0,0 @@
-run:
-  timeout: 5m
-linters:
-  enable:
-    - asasalint
-    - bidichk
-    - bodyclose
-    - containedctx
-    - contextcheck
-    - exportloopref
-    - gocheckcompilerdirectives
-    # conditionally enable this on linux/macos
-    # - gofmt
-    # - goimports
-    - intrange
-    - misspell
-    - nilerr
-    - nolintlint
-    - nosprintfhostport
-    - testifylint
-    - unconvert
-    - unused
-    - wastedassign
-    - whitespace
-    - usestdlibvars
-severity:
-  default-severity: error
-  rules:
-    - linters:
-        - gofmt
-        - goimports
-        - intrange
-        - usestdlibvars
-      severity: info

Dockerfile

@@ -1,144 +1,15 @@
-ARG GOLANG_VERSION=1.22.5
-ARG CMAKE_VERSION=3.22.1
-# this CUDA_VERSION corresponds with the one specified in docs/gpu.md
-ARG CUDA_VERSION=11.3.1
-ARG ROCM_VERSION=6.1.2
-
-# Copy the minimal context we need to run the generate scripts
-FROM scratch AS llm-code
-COPY .git .git
-COPY .gitmodules .gitmodules
-COPY llm llm
-
-FROM --platform=linux/amd64 nvidia/cuda:$CUDA_VERSION-devel-centos7 AS cuda-build-amd64
-ARG CMAKE_VERSION
-COPY ./scripts/rh_linux_deps.sh /
-RUN CMAKE_VERSION=${CMAKE_VERSION} sh /rh_linux_deps.sh
-ENV PATH /opt/rh/devtoolset-10/root/usr/bin:$PATH
-COPY --from=llm-code / /go/src/github.com/ollama/ollama/
-WORKDIR /go/src/github.com/ollama/ollama/llm/generate
-ARG CGO_CFLAGS
-RUN OLLAMA_SKIP_STATIC_GENERATE=1 OLLAMA_SKIP_CPU_GENERATE=1 sh gen_linux.sh
-
-FROM --platform=linux/arm64 nvidia/cuda:$CUDA_VERSION-devel-rockylinux8 AS cuda-build-arm64
-ARG CMAKE_VERSION
-COPY ./scripts/rh_linux_deps.sh /
-RUN CMAKE_VERSION=${CMAKE_VERSION} sh /rh_linux_deps.sh
-ENV PATH /opt/rh/gcc-toolset-10/root/usr/bin:$PATH
-COPY --from=llm-code / /go/src/github.com/ollama/ollama/
-WORKDIR /go/src/github.com/ollama/ollama/llm/generate
-ARG CGO_CFLAGS
-RUN OLLAMA_SKIP_STATIC_GENERATE=1 OLLAMA_SKIP_CPU_GENERATE=1 sh gen_linux.sh
-
-FROM --platform=linux/amd64 rocm/dev-centos-7:${ROCM_VERSION}-complete AS rocm-build-amd64
-ARG CMAKE_VERSION
-COPY ./scripts/rh_linux_deps.sh /
-RUN CMAKE_VERSION=${CMAKE_VERSION} sh /rh_linux_deps.sh
-ENV PATH /opt/rh/devtoolset-10/root/usr/bin:$PATH
-ENV LIBRARY_PATH /opt/amdgpu/lib64
-COPY --from=llm-code / /go/src/github.com/ollama/ollama/
-WORKDIR /go/src/github.com/ollama/ollama/llm/generate
-ARG CGO_CFLAGS
-ARG AMDGPU_TARGETS
-RUN OLLAMA_SKIP_STATIC_GENERATE=1 OLLAMA_SKIP_CPU_GENERATE=1 sh gen_linux.sh
-RUN mkdir /tmp/scratch && \
-    for dep in $(zcat /go/src/github.com/ollama/ollama/llm/build/linux/x86_64/rocm*/bin/deps.txt.gz) ; do \
-        cp ${dep} /tmp/scratch/ || exit 1 ; \
-    done && \
-    (cd /opt/rocm/lib && tar cf - rocblas/library) | (cd /tmp/scratch/ && tar xf - ) && \
-    mkdir -p /go/src/github.com/ollama/ollama/dist/deps/ && \
-    (cd /tmp/scratch/ && tar czvf /go/src/github.com/ollama/ollama/dist/deps/ollama-linux-amd64-rocm.tgz . )
-
-
-FROM --platform=linux/amd64 centos:7 AS cpu-builder-amd64
-ARG CMAKE_VERSION
-ARG GOLANG_VERSION
-COPY ./scripts/rh_linux_deps.sh /
-RUN CMAKE_VERSION=${CMAKE_VERSION} GOLANG_VERSION=${GOLANG_VERSION} sh /rh_linux_deps.sh
-ENV PATH /opt/rh/devtoolset-10/root/usr/bin:$PATH
-COPY --from=llm-code / /go/src/github.com/ollama/ollama/
-ARG OLLAMA_CUSTOM_CPU_DEFS
-ARG CGO_CFLAGS
-WORKDIR /go/src/github.com/ollama/ollama/llm/generate
-
-FROM --platform=linux/amd64 cpu-builder-amd64 AS static-build-amd64
-RUN OLLAMA_CPU_TARGET="static" sh gen_linux.sh
-FROM --platform=linux/amd64 cpu-builder-amd64 AS cpu-build-amd64
-RUN OLLAMA_SKIP_STATIC_GENERATE=1 OLLAMA_CPU_TARGET="cpu" sh gen_linux.sh
-FROM --platform=linux/amd64 cpu-builder-amd64 AS cpu_avx-build-amd64
-RUN OLLAMA_SKIP_STATIC_GENERATE=1 OLLAMA_CPU_TARGET="cpu_avx" sh gen_linux.sh
-FROM --platform=linux/amd64 cpu-builder-amd64 AS cpu_avx2-build-amd64
-RUN OLLAMA_SKIP_STATIC_GENERATE=1 OLLAMA_CPU_TARGET="cpu_avx2" sh gen_linux.sh
-
-FROM --platform=linux/arm64 rockylinux:8 AS cpu-builder-arm64
-ARG CMAKE_VERSION
-ARG GOLANG_VERSION
-COPY ./scripts/rh_linux_deps.sh /
-RUN CMAKE_VERSION=${CMAKE_VERSION} GOLANG_VERSION=${GOLANG_VERSION} sh /rh_linux_deps.sh
-ENV PATH /opt/rh/gcc-toolset-10/root/usr/bin:$PATH
-COPY --from=llm-code / /go/src/github.com/ollama/ollama/
-ARG OLLAMA_CUSTOM_CPU_DEFS
-ARG CGO_CFLAGS
-WORKDIR /go/src/github.com/ollama/ollama/llm/generate
-
-FROM --platform=linux/arm64 cpu-builder-arm64 AS static-build-arm64
-RUN OLLAMA_CPU_TARGET="static" sh gen_linux.sh
-FROM --platform=linux/arm64 cpu-builder-arm64 AS cpu-build-arm64
-RUN OLLAMA_SKIP_STATIC_GENERATE=1 OLLAMA_CPU_TARGET="cpu" sh gen_linux.sh
-
-
-# Intermediate stage used for ./scripts/build_linux.sh
-FROM --platform=linux/amd64 cpu-build-amd64 AS build-amd64
-ENV CGO_ENABLED 1
-WORKDIR /go/src/github.com/ollama/ollama
+FROM golang:1.20
+WORKDIR /go/src/github.com/jmorganca/ollama
 COPY . .
-COPY --from=static-build-amd64 /go/src/github.com/ollama/ollama/llm/build/linux/ llm/build/linux/
-COPY --from=cpu_avx-build-amd64 /go/src/github.com/ollama/ollama/llm/build/linux/ llm/build/linux/
-COPY --from=cpu_avx2-build-amd64 /go/src/github.com/ollama/ollama/llm/build/linux/ llm/build/linux/
-COPY --from=cuda-build-amd64 /go/src/github.com/ollama/ollama/llm/build/linux/ llm/build/linux/
-COPY --from=rocm-build-amd64 /go/src/github.com/ollama/ollama/llm/build/linux/ llm/build/linux/
-COPY --from=rocm-build-amd64 /go/src/github.com/ollama/ollama/dist/deps/ ./dist/deps/
-ARG GOFLAGS
-ARG CGO_CFLAGS
-RUN go build -trimpath .
+RUN CGO_ENABLED=1 go build -ldflags '-linkmode external -extldflags "-static"' .
 
-# Intermediate stage used for ./scripts/build_linux.sh
-FROM --platform=linux/arm64 cpu-build-arm64 AS build-arm64
-ENV CGO_ENABLED 1
-ARG GOLANG_VERSION
-WORKDIR /go/src/github.com/ollama/ollama
-COPY . .
-COPY --from=static-build-arm64 /go/src/github.com/ollama/ollama/llm/build/linux/ llm/build/linux/
-COPY --from=cuda-build-arm64 /go/src/github.com/ollama/ollama/llm/build/linux/ llm/build/linux/
-ARG GOFLAGS
-ARG CGO_CFLAGS
-RUN go build -trimpath .
-
-# Runtime stages
-FROM --platform=linux/amd64 ubuntu:22.04 as runtime-amd64
-RUN apt-get update && apt-get install -y ca-certificates
-COPY --from=build-amd64 /go/src/github.com/ollama/ollama/ollama /bin/ollama
-FROM --platform=linux/arm64 ubuntu:22.04 as runtime-arm64
-RUN apt-get update && apt-get install -y ca-certificates
-COPY --from=build-arm64 /go/src/github.com/ollama/ollama/ollama /bin/ollama
-
-# Radeon images are much larger so we keep it distinct from the CPU/CUDA image
-FROM --platform=linux/amd64 rocm/dev-centos-7:${ROCM_VERSION}-complete as runtime-rocm
-RUN update-pciids
-COPY --from=build-amd64 /go/src/github.com/ollama/ollama/ollama /bin/ollama
+FROM alpine
+COPY --from=0 /go/src/github.com/jmorganca/ollama/ollama /bin/ollama
 EXPOSE 11434
-ENV OLLAMA_HOST 0.0.0.0
-
+ARG USER=ollama
+ARG GROUP=ollama
+RUN addgroup -g 1000 $GROUP && adduser -u 1000 -DG $GROUP $USER
+USER $USER:$GROUP
 ENTRYPOINT ["/bin/ollama"]
-CMD ["serve"]
-
-FROM runtime-$TARGETARCH
-EXPOSE 11434
 ENV OLLAMA_HOST 0.0.0.0
-ENV PATH=/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin
-ENV LD_LIBRARY_PATH=/usr/local/nvidia/lib:/usr/local/nvidia/lib64
-ENV NVIDIA_DRIVER_CAPABILITIES=compute,utility
-ENV NVIDIA_VISIBLE_DEVICES=all
-
-ENTRYPOINT ["/bin/ollama"]
 CMD ["serve"]

README.md

@@ -1,118 +1,74 @@
 <div align="center">
- <img alt="ollama" height="200px" src="https://github.com/ollama/ollama/assets/3325447/0d0b44e2-8f4a-4e99-9b52-a5c1c741c8f7">
+  <picture>
+    <source media="(prefers-color-scheme: dark)" height="200px" srcset="https://github.com/jmorganca/ollama/assets/3325447/56ea1849-1284-4645-8970-956de6e51c3c">
+    <img alt="logo" height="200px" src="https://github.com/jmorganca/ollama/assets/3325447/0d0b44e2-8f4a-4e99-9b52-a5c1c741c8f7">
+  </picture>
 </div>
 
 # Ollama
 
 [![Discord](https://dcbadge.vercel.app/api/server/ollama?style=flat&compact=true)](https://discord.gg/ollama)
 
-Get up and running with large language models.
+> Note: Ollama is in early preview. Please report any issues you find.
 
-### macOS
+Run, create, and share large language models (LLMs).
 
-[Download](https://ollama.com/download/Ollama-darwin.zip)
+## Download
 
-### Windows preview
-
-[Download](https://ollama.com/download/OllamaSetup.exe)
-
-### Linux
-
-```
-curl -fsSL https://ollama.com/install.sh | sh
-```
-
-[Manual install instructions](https://github.com/ollama/ollama/blob/main/docs/linux.md)
-
-### Docker
-
-The official [Ollama Docker image](https://hub.docker.com/r/ollama/ollama) `ollama/ollama` is available on Docker Hub.
-
-### Libraries
-
-- [ollama-python](https://github.com/ollama/ollama-python)
-- [ollama-js](https://github.com/ollama/ollama-js)
+- [Download](https://ollama.ai/download) for macOS on Apple Silicon (Intel coming soon)
+- Download for Windows and Linux (coming soon)
+- Build [from source](#building)
 
 ## Quickstart
 
-To run and chat with [Llama 3.1](https://ollama.com/library/llama3.1):
+To run and chat with [Llama 2](https://ai.meta.com/llama), the new model by Meta:
 
 ```
-ollama run llama3.1
+ollama run llama2
 ```
 
 ## Model library
 
-Ollama supports a list of models available on [ollama.com/library](https://ollama.com/library 'ollama model library')
+`ollama` includes a library of open-source models:
 
-Here are some example models that can be downloaded:
+| Model                    | Parameters | Size  | Download                    |
+| ------------------------ | ---------- | ----- | --------------------------- |
+| Llama2                   | 7B         | 3.8GB | `ollama pull llama2`        |
+| Llama2 13B               | 13B        | 7.3GB | `ollama pull llama2:13b`    |
+| Orca Mini                | 3B         | 1.9GB | `ollama pull orca`          |
+| Vicuna                   | 7B         | 3.8GB | `ollama pull vicuna`        |
+| Nous-Hermes              | 13B        | 7.3GB | `ollama pull nous-hermes`   |
+| Wizard Vicuna Uncensored | 13B        | 7.3GB | `ollama pull wizard-vicuna` |
 
-| Model              | Parameters | Size  | Download                       |
-| ------------------ | ---------- | ----- | ------------------------------ |
-| Llama 3.1          | 8B         | 4.7GB | `ollama run llama3.1`          |
-| Llama 3.1          | 70B        | 40GB  | `ollama run llama3.1:70b`      |
-| Llama 3.1          | 405B       | 231GB | `ollama run llama3.1:405b`     |
-| Phi 3 Mini         | 3.8B       | 2.3GB | `ollama run phi3`              |
-| Phi 3 Medium       | 14B        | 7.9GB | `ollama run phi3:medium`       |
-| Gemma 2            | 9B         | 5.5GB | `ollama run gemma2`            |
-| Gemma 2            | 27B        | 16GB  | `ollama run gemma2:27b`        |
-| Mistral            | 7B         | 4.1GB | `ollama run mistral`           |
-| Moondream 2        | 1.4B       | 829MB | `ollama run moondream`         |
-| Neural Chat        | 7B         | 4.1GB | `ollama run neural-chat`       |
-| Starling           | 7B         | 4.1GB | `ollama run starling-lm`       |
-| Code Llama         | 7B         | 3.8GB | `ollama run codellama`         |
-| Llama 2 Uncensored | 7B         | 3.8GB | `ollama run llama2-uncensored` |
-| LLaVA              | 7B         | 4.5GB | `ollama run llava`             |
-| Solar              | 10.7B      | 6.1GB | `ollama run solar`             |
+> Note: You should have at least 8 GB of RAM to run the 3B models, 16 GB to run the 7B models, and 32 GB to run the 13B models.
 
-> [!NOTE]
-> You should have at least 8 GB of RAM available to run the 7B models, 16 GB to run the 13B models, and 32 GB to run the 33B models.
+## Examples
 
-## Customize a model
-
-### Import from GGUF
-
-Ollama supports importing GGUF models in the Modelfile:
-
-1. Create a file named `Modelfile`, with a `FROM` instruction with the local filepath to the model you want to import.
-
-   ```
-   FROM ./vicuna-33b.Q4_0.gguf
-   ```
-
-2. Create the model in Ollama
-
-   ```
-   ollama create example -f Modelfile
-   ```
-
-3. Run the model
-
-   ```
-   ollama run example
-   ```
-
-### Import from PyTorch or Safetensors
-
-See the [guide](docs/import.md) on importing models for more information.
-
-### Customize a prompt
-
-Models from the Ollama library can be customized with a prompt. For example, to customize the `llama3.1` model:
+### Run a model
 
 ```
-ollama pull llama3.1
+ollama run llama2
+>>> hi
+Hello! How can I help you today?
+```
+
+### Create a custom model
+
+Pull a base model:
+
+```
+ollama pull llama2
 ```
 
 Create a `Modelfile`:
 
 ```
-FROM llama3.1
+FROM llama2
 
 # set the temperature to 1 [higher is more creative, lower is more coherent]
 PARAMETER temperature 1
 
-# set the system message
+# set the system prompt
 SYSTEM """
 You are Mario from Super Mario Bros. Answer as Mario, the assistant, only.
 """
@@ -127,280 +83,55 @@ ollama run mario
 Hello! It's your friend Mario.
 ```
 
-For more examples, see the [examples](examples) directory. For more information on working with a Modelfile, see the [Modelfile](docs/modelfile.md) documentation.
+For more examples, see the [examples](./examples) directory.
 
-## CLI Reference
-
-### Create a model
-
-`ollama create` is used to create a model from a Modelfile.
+### Pull a model from the registry
 
 ```
-ollama create mymodel -f ./Modelfile
+ollama pull orca
 ```
 
-### Pull a model
-
-```
-ollama pull llama3.1
-```
-
-> This command can also be used to update a local model. Only the diff will be pulled.
-
-### Remove a model
-
-```
-ollama rm llama3.1
-```
-
-### Copy a model
-
-```
-ollama cp llama3.1 my-model
-```
-
-### Multiline input
-
-For multiline input, you can wrap text with `"""`:
-
-```
->>> """Hello,
-... world!
-... """
-I'm a basic program that prints the famous "Hello, world!" message to the console.
-```
-
-### Multimodal models
-
-```
-ollama run llava "What's in this image? /Users/jmorgan/Desktop/smile.png"
-The image features a yellow smiley face, which is likely the central focus of the picture.
-```
-
-### Pass the prompt as an argument
-
-```
-$ ollama run llama3.1 "Summarize this file: $(cat README.md)"
- Ollama is a lightweight, extensible framework for building and running language models on the local machine. It provides a simple API for creating, running, and managing models, as well as a library of pre-built models that can be easily used in a variety of applications.
-```
-
-### Show model information
-
-```
-ollama show llama3.1
-```
-
-### List models on your computer
+### Listing local models
 
 ```
 ollama list
 ```
 
-### Start Ollama
+## Model packages
 
-`ollama serve` is used when you want to start ollama without running the desktop application.
+### Overview
+
+Ollama bundles model weights, configuration, and data into a single package, defined by a [Modelfile](./docs/modelfile.md).
+
+<picture>
+  <source media="(prefers-color-scheme: dark)" height="480" srcset="https://github.com/jmorganca/ollama/assets/251292/2fd96b5f-191b-45c1-9668-941cfad4eb70">
+  <img alt="logo" height="480" src="https://github.com/jmorganca/ollama/assets/251292/2fd96b5f-191b-45c1-9668-941cfad4eb70">
+</picture>
 
 ## Building
 
-See the [developer guide](https://github.com/ollama/ollama/blob/main/docs/development.md)
-
-### Running local builds
-
-Next, start the server:
-
 ```
-./ollama serve
+go build .
 ```
 
-Finally, in a separate shell, run a model:
+To run it start the server:
 
 ```
-./ollama run llama3.1
+./ollama serve &
+```
+
+Finally, run a model!
+
+```
+./ollama run llama2
 ```
 
 ## REST API
 
-Ollama has a REST API for running and managing models.
+### `POST /api/generate`
 
-### Generate a response
+Generate text from a model.
 
 ```
-curl http://localhost:11434/api/generate -d '{
-  "model": "llama3.1",
-  "prompt":"Why is the sky blue?"
-}'
+curl -X POST http://localhost:11434/api/generate -d '{"model": "llama2", "prompt":"Why is the sky blue?"}'
 ```
-
-### Chat with a model
-
-```
-curl http://localhost:11434/api/chat -d '{
-  "model": "llama3.1",
-  "messages": [
-    { "role": "user", "content": "why is the sky blue?" }
-  ]
-}'
-```
-
-See the [API documentation](./docs/api.md) for all endpoints.
-
-## Community Integrations
-
-### Web & Desktop
-
-- [Open WebUI](https://github.com/open-webui/open-webui)
-- [Enchanted (macOS native)](https://github.com/AugustDev/enchanted)
-- [Hollama](https://github.com/fmaclen/hollama)
-- [Lollms-Webui](https://github.com/ParisNeo/lollms-webui)
-- [LibreChat](https://github.com/danny-avila/LibreChat)
-- [Bionic GPT](https://github.com/bionic-gpt/bionic-gpt)
-- [HTML UI](https://github.com/rtcfirefly/ollama-ui)
-- [Saddle](https://github.com/jikkuatwork/saddle)
-- [Chatbot UI](https://github.com/ivanfioravanti/chatbot-ollama)
-- [Chatbot UI v2](https://github.com/mckaywrigley/chatbot-ui)
-- [Typescript UI](https://github.com/ollama-interface/Ollama-Gui?tab=readme-ov-file)
-- [Minimalistic React UI for Ollama Models](https://github.com/richawo/minimal-llm-ui)
-- [Ollamac](https://github.com/kevinhermawan/Ollamac)
-- [big-AGI](https://github.com/enricoros/big-AGI/blob/main/docs/config-local-ollama.md)
-- [Cheshire Cat assistant framework](https://github.com/cheshire-cat-ai/core)
-- [Amica](https://github.com/semperai/amica)
-- [chatd](https://github.com/BruceMacD/chatd)
-- [Ollama-SwiftUI](https://github.com/kghandour/Ollama-SwiftUI)
-- [Dify.AI](https://github.com/langgenius/dify)
-- [MindMac](https://mindmac.app)
-- [NextJS Web Interface for Ollama](https://github.com/jakobhoeg/nextjs-ollama-llm-ui)
-- [Msty](https://msty.app)
-- [Chatbox](https://github.com/Bin-Huang/Chatbox)
-- [WinForm Ollama Copilot](https://github.com/tgraupmann/WinForm_Ollama_Copilot)
-- [NextChat](https://github.com/ChatGPTNextWeb/ChatGPT-Next-Web) with [Get Started Doc](https://docs.nextchat.dev/models/ollama)
-- [Alpaca WebUI](https://github.com/mmo80/alpaca-webui)
-- [OllamaGUI](https://github.com/enoch1118/ollamaGUI)
-- [OpenAOE](https://github.com/InternLM/OpenAOE)
-- [Odin Runes](https://github.com/leonid20000/OdinRunes)
-- [LLM-X](https://github.com/mrdjohnson/llm-x) (Progressive Web App)
-- [AnythingLLM (Docker + MacOs/Windows/Linux native app)](https://github.com/Mintplex-Labs/anything-llm)
-- [Ollama Basic Chat: Uses HyperDiv Reactive UI](https://github.com/rapidarchitect/ollama_basic_chat)
-- [Ollama-chats RPG](https://github.com/drazdra/ollama-chats)
-- [QA-Pilot](https://github.com/reid41/QA-Pilot) (Chat with Code Repository)
-- [ChatOllama](https://github.com/sugarforever/chat-ollama) (Open Source Chatbot based on Ollama with Knowledge Bases)
-- [CRAG Ollama Chat](https://github.com/Nagi-ovo/CRAG-Ollama-Chat) (Simple Web Search with Corrective RAG)
-- [RAGFlow](https://github.com/infiniflow/ragflow) (Open-source Retrieval-Augmented Generation engine based on deep document understanding)
-- [StreamDeploy](https://github.com/StreamDeploy-DevRel/streamdeploy-llm-app-scaffold) (LLM Application Scaffold)
-- [chat](https://github.com/swuecho/chat) (chat web app for teams)
-- [Lobe Chat](https://github.com/lobehub/lobe-chat) with [Integrating Doc](https://lobehub.com/docs/self-hosting/examples/ollama)
-- [Ollama RAG Chatbot](https://github.com/datvodinh/rag-chatbot.git) (Local Chat with multiple PDFs using Ollama and RAG)
-- [BrainSoup](https://www.nurgo-software.com/products/brainsoup) (Flexible native client with RAG & multi-agent automation)
-- [macai](https://github.com/Renset/macai) (macOS client for Ollama, ChatGPT, and other compatible API back-ends)
-- [Olpaka](https://github.com/Otacon/olpaka) (User-friendly Flutter Web App for Ollama)
-- [OllamaSpring](https://github.com/CrazyNeil/OllamaSpring) (Ollama Client for macOS)
-- [LLocal.in](https://github.com/kartikm7/llocal) (Easy to use Electron Desktop Client for Ollama)
-- [Ollama with Google Mesop](https://github.com/rapidarchitect/ollama_mesop/) (Mesop Chat Client implementation with Ollama)
-- [Kerlig AI](https://www.kerlig.com/) (AI writing assistant for macOS)
-- [AI Studio](https://github.com/MindWorkAI/AI-Studio)
-- [Sidellama](https://github.com/gyopak/sidellama) (browser-based LLM client)
-- [LLMStack](https://github.com/trypromptly/LLMStack) (No-code multi-agent framework to build LLM agents and workflows)
-- [BoltAI for Mac](https://boltai.com) (AI Chat Client for Mac)
-
-### Terminal
-
-- [oterm](https://github.com/ggozad/oterm)
-- [Ellama Emacs client](https://github.com/s-kostyaev/ellama)
-- [Emacs client](https://github.com/zweifisch/ollama)
-- [gen.nvim](https://github.com/David-Kunz/gen.nvim)
-- [ollama.nvim](https://github.com/nomnivore/ollama.nvim)
-- [ollero.nvim](https://github.com/marco-souza/ollero.nvim)
-- [ollama-chat.nvim](https://github.com/gerazov/ollama-chat.nvim)
-- [ogpt.nvim](https://github.com/huynle/ogpt.nvim)
-- [gptel Emacs client](https://github.com/karthink/gptel)
-- [Oatmeal](https://github.com/dustinblackman/oatmeal)
-- [cmdh](https://github.com/pgibler/cmdh)
-- [ooo](https://github.com/npahlfer/ooo)
-- [shell-pilot](https://github.com/reid41/shell-pilot)
-- [tenere](https://github.com/pythops/tenere)
-- [llm-ollama](https://github.com/taketwo/llm-ollama) for [Datasette's LLM CLI](https://llm.datasette.io/en/stable/).
-- [typechat-cli](https://github.com/anaisbetts/typechat-cli)
-- [ShellOracle](https://github.com/djcopley/ShellOracle)
-- [tlm](https://github.com/yusufcanb/tlm)
-- [podman-ollama](https://github.com/ericcurtin/podman-ollama)
-- [gollama](https://github.com/sammcj/gollama)
-
-### Database
-
-- [MindsDB](https://github.com/mindsdb/mindsdb/blob/staging/mindsdb/integrations/handlers/ollama_handler/README.md) (Connects Ollama models with nearly 200 data platforms and apps)
-- [chromem-go](https://github.com/philippgille/chromem-go/blob/v0.5.0/embed_ollama.go) with [example](https://github.com/philippgille/chromem-go/tree/v0.5.0/examples/rag-wikipedia-ollama)
-
-### Package managers
-
-- [Pacman](https://archlinux.org/packages/extra/x86_64/ollama/)
-- [Helm Chart](https://artifacthub.io/packages/helm/ollama-helm/ollama)
-- [Guix channel](https://codeberg.org/tusharhero/ollama-guix)
-
-### Libraries
-
-- [LangChain](https://python.langchain.com/docs/integrations/llms/ollama) and [LangChain.js](https://js.langchain.com/docs/modules/model_io/models/llms/integrations/ollama) with [example](https://js.langchain.com/docs/use_cases/question_answering/local_retrieval_qa)
-- [Firebase Genkit](https://firebase.google.com/docs/genkit/plugins/ollama)
-- [LangChainGo](https://github.com/tmc/langchaingo/) with [example](https://github.com/tmc/langchaingo/tree/main/examples/ollama-completion-example)
-- [LangChain4j](https://github.com/langchain4j/langchain4j) with [example](https://github.com/langchain4j/langchain4j-examples/tree/main/ollama-examples/src/main/java)
-- [LangChainRust](https://github.com/Abraxas-365/langchain-rust) with [example](https://github.com/Abraxas-365/langchain-rust/blob/main/examples/llm_ollama.rs)
-- [LlamaIndex](https://gpt-index.readthedocs.io/en/stable/examples/llm/ollama.html)
-- [LiteLLM](https://github.com/BerriAI/litellm)
-- [OllamaSharp for .NET](https://github.com/awaescher/OllamaSharp)
-- [Ollama for Ruby](https://github.com/gbaptista/ollama-ai)
-- [Ollama-rs for Rust](https://github.com/pepperoni21/ollama-rs)
-- [Ollama-hpp for C++](https://github.com/jmont-dev/ollama-hpp)
-- [Ollama4j for Java](https://github.com/amithkoujalgi/ollama4j)
-- [ModelFusion Typescript Library](https://modelfusion.dev/integration/model-provider/ollama)
-- [OllamaKit for Swift](https://github.com/kevinhermawan/OllamaKit)
-- [Ollama for Dart](https://github.com/breitburg/dart-ollama)
-- [Ollama for Laravel](https://github.com/cloudstudio/ollama-laravel)
-- [LangChainDart](https://github.com/davidmigloz/langchain_dart)
-- [Semantic Kernel - Python](https://github.com/microsoft/semantic-kernel/tree/main/python/semantic_kernel/connectors/ai/ollama)
-- [Haystack](https://github.com/deepset-ai/haystack-integrations/blob/main/integrations/ollama.md)
-- [Elixir LangChain](https://github.com/brainlid/langchain)
-- [Ollama for R - rollama](https://github.com/JBGruber/rollama)
-- [Ollama for R - ollama-r](https://github.com/hauselin/ollama-r)
-- [Ollama-ex for Elixir](https://github.com/lebrunel/ollama-ex)
-- [Ollama Connector for SAP ABAP](https://github.com/b-tocs/abap_btocs_ollama)
-- [Testcontainers](https://testcontainers.com/modules/ollama/)
-- [Portkey](https://portkey.ai/docs/welcome/integration-guides/ollama)
-- [PromptingTools.jl](https://github.com/svilupp/PromptingTools.jl) with an [example](https://svilupp.github.io/PromptingTools.jl/dev/examples/working_with_ollama)
-- [LlamaScript](https://github.com/Project-Llama/llamascript)
-
-### Mobile
-
-- [Enchanted](https://github.com/AugustDev/enchanted)
-- [Maid](https://github.com/Mobile-Artificial-Intelligence/maid)
-
-### Extensions & Plugins
-
-- [Raycast extension](https://github.com/MassimilianoPasquini97/raycast_ollama)
-- [Discollama](https://github.com/mxyng/discollama) (Discord bot inside the Ollama discord channel)
-- [Continue](https://github.com/continuedev/continue)
-- [Obsidian Ollama plugin](https://github.com/hinterdupfinger/obsidian-ollama)
-- [Logseq Ollama plugin](https://github.com/omagdy7/ollama-logseq)
-- [NotesOllama](https://github.com/andersrex/notesollama) (Apple Notes Ollama plugin)
-- [Dagger Chatbot](https://github.com/samalba/dagger-chatbot)
-- [Discord AI Bot](https://github.com/mekb-turtle/discord-ai-bot)
-- [Ollama Telegram Bot](https://github.com/ruecat/ollama-telegram)
-- [Hass Ollama Conversation](https://github.com/ej52/hass-ollama-conversation)
-- [Rivet plugin](https://github.com/abrenneke/rivet-plugin-ollama)
-- [Obsidian BMO Chatbot plugin](https://github.com/longy2k/obsidian-bmo-chatbot)
-- [Cliobot](https://github.com/herval/cliobot) (Telegram bot with Ollama support)
-- [Copilot for Obsidian plugin](https://github.com/logancyang/obsidian-copilot)
-- [Obsidian Local GPT plugin](https://github.com/pfrankov/obsidian-local-gpt)
-- [Open Interpreter](https://docs.openinterpreter.com/language-model-setup/local-models/ollama)
-- [Llama Coder](https://github.com/ex3ndr/llama-coder) (Copilot alternative using Ollama)
-- [Ollama Copilot](https://github.com/bernardo-bruning/ollama-copilot) (Proxy that allows you to use ollama as a copilot like Github copilot)
-- [twinny](https://github.com/rjmacarthy/twinny) (Copilot and Copilot chat alternative using Ollama)
-- [Wingman-AI](https://github.com/RussellCanfield/wingman-ai) (Copilot code and chat alternative using Ollama and Hugging Face)
-- [Page Assist](https://github.com/n4ze3m/page-assist) (Chrome Extension)
-- [AI Telegram Bot](https://github.com/tusharhero/aitelegrambot) (Telegram bot using Ollama in backend)
-- [AI ST Completion](https://github.com/yaroslavyaroslav/OpenAI-sublime-text) (Sublime Text 4 AI assistant plugin with Ollama support)
-- [Discord-Ollama Chat Bot](https://github.com/kevinthedang/discord-ollama) (Generalized TypeScript Discord Bot w/ Tuning Documentation)
-- [Discord AI chat/moderation bot](https://github.com/rapmd73/Companion) Chat/moderation bot written in python. Uses Ollama to create personalities.
-- [Headless Ollama](https://github.com/nischalj10/headless-ollama) (Scripts to automatically install ollama client & models on any OS for apps that depends on ollama server)
-
-### Supported backends
-
-- [llama.cpp](https://github.com/ggerganov/llama.cpp) project founded by Georgi Gerganov.
-

SECURITY.md

@@ -1,25 +0,0 @@
-# Security
-
-The Ollama maintainer team takes security seriously and will actively work to resolve security issues.
-
-## Reporting a vulnerability
-
-If you discover a security vulnerability, please do not open a public issue. Instead, please report it by emailing hello@ollama.com. We ask that you give us sufficient time to investigate and address the vulnerability before disclosing it publicly.
-
-Please include the following details in your report:
-- A description of the vulnerability
-- Steps to reproduce the issue
-- Your assessment of the potential impact
-- Any possible mitigations
-
-## Security best practices
-
-While the maintainer team does their best to secure Ollama, users are encouraged to implement their own security best practices, such as:
-
-- Regularly updating to the latest version of Ollama
-- Securing access to hosted instances of Ollama
-- Monitoring systems for unusual activity
-
-## Contact
-
-For any other questions or concerns related to security, please contact us at hello@ollama.com

api/client.go

@@ -1,16 +1,3 @@
-// Package api implements the client-side API for code wishing to interact
-// with the ollama service. The methods of the [Client] type correspond to
-// the ollama REST API as described in [the API documentation].
-// The ollama command-line client itself uses this package to interact with
-// the backend service.
-//
-// # Examples
-//
-// Several examples of using this package are available [in the GitHub
-// repository].
-//
-// [the API documentation]: https://github.com/ollama/ollama/blob/main/docs/api.md
-// [in the GitHub repository]: https://github.com/ollama/ollama/tree/main/examples
 package api
 
 import (
@@ -22,22 +9,16 @@ import (
 	"io"
 	"net/http"
 	"net/url"
-	"runtime"
-
-	"github.com/ollama/ollama/envconfig"
-	"github.com/ollama/ollama/format"
-	"github.com/ollama/ollama/version"
 )
 
-// Client encapsulates client state for interacting with the ollama
-// service. Use [ClientFromEnvironment] to create new Clients.
 type Client struct {
-	base *url.URL
-	http *http.Client
+	base    url.URL
+	HTTP    http.Client
+	Headers http.Header
 }
 
 func checkError(resp *http.Response, body []byte) error {
-	if resp.StatusCode < http.StatusBadRequest {
+	if resp.StatusCode >= 200 && resp.StatusCode < 400 {
 		return nil
 	}
 
@@ -52,26 +33,15 @@ func checkError(resp *http.Response, body []byte) error {
 	return apiError
 }
 
-// ClientFromEnvironment creates a new [Client] using configuration from the
-// environment variable OLLAMA_HOST, which points to the network host and
-// port on which the ollama service is listenting. The format of this variable
-// is:
-//
-//	<scheme>://<host>:<port>
-//
-// If the variable is not specified, a default ollama host and port will be
-// used.
-func ClientFromEnvironment() (*Client, error) {
-	return &Client{
-		base: envconfig.Host(),
-		http: http.DefaultClient,
-	}, nil
-}
+func NewClient(hosts ...string) *Client {
+	host := "127.0.0.1:11434"
+	if len(hosts) > 0 {
+		host = hosts[0]
+	}
 
-func NewClient(base *url.URL, http *http.Client) *Client {
 	return &Client{
-		base: base,
-		http: http,
+		base: url.URL{Scheme: "http", Host: host},
+		HTTP: http.Client{},
 	}
 }
 
@@ -79,33 +49,29 @@ func (c *Client) do(ctx context.Context, method, path string, reqData, respData
 	var reqBody io.Reader
 	var data []byte
 	var err error
-
-	switch reqData := reqData.(type) {
-	case io.Reader:
-		// reqData is already an io.Reader
-		reqBody = reqData
-	case nil:
-		// noop
-	default:
+	if reqData != nil {
 		data, err = json.Marshal(reqData)
 		if err != nil {
 			return err
 		}
-
 		reqBody = bytes.NewReader(data)
 	}
 
-	requestURL := c.base.JoinPath(path)
-	request, err := http.NewRequestWithContext(ctx, method, requestURL.String(), reqBody)
+	url := c.base.JoinPath(path).String()
+
+	req, err := http.NewRequestWithContext(ctx, method, url, reqBody)
 	if err != nil {
 		return err
 	}
 
-	request.Header.Set("Content-Type", "application/json")
-	request.Header.Set("Accept", "application/json")
-	request.Header.Set("User-Agent", fmt.Sprintf("ollama/%s (%s %s) Go/%s", version.Version, runtime.GOARCH, runtime.GOOS, runtime.Version()))
+	req.Header.Set("Content-Type", "application/json")
+	req.Header.Set("Accept", "application/json")
 
-	respObj, err := c.http.Do(request)
+	for k, v := range c.Headers {
+		req.Header[k] = v
+	}
+
+	respObj, err := c.HTTP.Do(req)
 	if err != nil {
 		return err
 	}
@@ -128,8 +94,6 @@ func (c *Client) do(ctx context.Context, method, path string, reqData, respData
 	return nil
 }
 
-const maxBufferSize = 512 * format.KiloByte
-
 func (c *Client) stream(ctx context.Context, method, path string, data any, fn func([]byte) error) error {
 	var buf *bytes.Buffer
 	if data != nil {
@@ -141,26 +105,21 @@ func (c *Client) stream(ctx context.Context, method, path string, data any, fn f
 		buf = bytes.NewBuffer(bts)
 	}
 
-	requestURL := c.base.JoinPath(path)
-	request, err := http.NewRequestWithContext(ctx, method, requestURL.String(), buf)
+	request, err := http.NewRequestWithContext(ctx, method, c.base.JoinPath(path).String(), buf)
 	if err != nil {
 		return err
 	}
 
 	request.Header.Set("Content-Type", "application/json")
-	request.Header.Set("Accept", "application/x-ndjson")
-	request.Header.Set("User-Agent", fmt.Sprintf("ollama/%s (%s %s) Go/%s", version.Version, runtime.GOARCH, runtime.GOOS, runtime.Version()))
+	request.Header.Set("Accept", "application/json")
 
-	response, err := c.http.Do(request)
+	response, err := http.DefaultClient.Do(request)
 	if err != nil {
 		return err
 	}
 	defer response.Body.Close()
 
 	scanner := bufio.NewScanner(response.Body)
-	// increase the buffer size to avoid running out of space
-	scanBuf := make([]byte, 0, maxBufferSize)
-	scanner.Buffer(scanBuf, maxBufferSize)
 	for scanner.Scan() {
 		var errorResponse struct {
 			Error string `json:"error,omitempty"`
@@ -172,10 +131,10 @@ func (c *Client) stream(ctx context.Context, method, path string, data any, fn f
 		}
 
 		if errorResponse.Error != "" {
-			return fmt.Errorf(errorResponse.Error)
+			return fmt.Errorf("stream: %s", errorResponse.Error)
 		}
 
-		if response.StatusCode >= http.StatusBadRequest {
+		if response.StatusCode >= 400 {
 			return StatusError{
 				StatusCode:   response.StatusCode,
 				Status:       response.Status,
@@ -191,14 +150,8 @@ func (c *Client) stream(ctx context.Context, method, path string, data any, fn f
 	return nil
 }
 
-// GenerateResponseFunc is a function that [Client.Generate] invokes every time
-// a response is received from the service. If this function returns an error,
-// [Client.Generate] will stop generating and return this error.
 type GenerateResponseFunc func(GenerateResponse) error
 
-// Generate generates a response for a given prompt. The req parameter should
-// be populated with prompt details. fn is called for each response (there may
-// be multiple responses, e.g. in case streaming is enabled).
 func (c *Client) Generate(ctx context.Context, req *GenerateRequest, fn GenerateResponseFunc) error {
 	return c.stream(ctx, http.MethodPost, "/api/generate", req, func(bts []byte) error {
 		var resp GenerateResponse
@@ -210,34 +163,8 @@ func (c *Client) Generate(ctx context.Context, req *GenerateRequest, fn Generate
 	})
 }
 
-// ChatResponseFunc is a function that [Client.Chat] invokes every time
-// a response is received from the service. If this function returns an error,
-// [Client.Chat] will stop generating and return this error.
-type ChatResponseFunc func(ChatResponse) error
-
-// Chat generates the next message in a chat. [ChatRequest] may contain a
-// sequence of messages which can be used to maintain chat history with a model.
-// fn is called for each response (there may be multiple responses, e.g. if case
-// streaming is enabled).
-func (c *Client) Chat(ctx context.Context, req *ChatRequest, fn ChatResponseFunc) error {
-	return c.stream(ctx, http.MethodPost, "/api/chat", req, func(bts []byte) error {
-		var resp ChatResponse
-		if err := json.Unmarshal(bts, &resp); err != nil {
-			return err
-		}
-
-		return fn(resp)
-	})
-}
-
-// PullProgressFunc is a function that [Client.Pull] invokes every time there
-// is progress with a "pull" request sent to the service. If this function
-// returns an error, [Client.Pull] will stop the process and return this error.
 type PullProgressFunc func(ProgressResponse) error
 
-// Pull downloads a model from the ollama library. fn is called each time
-// progress is made on the request and can be used to display a progress bar,
-// etc.
 func (c *Client) Pull(ctx context.Context, req *PullRequest, fn PullProgressFunc) error {
 	return c.stream(ctx, http.MethodPost, "/api/pull", req, func(bts []byte) error {
 		var resp ProgressResponse
@@ -249,14 +176,8 @@ func (c *Client) Pull(ctx context.Context, req *PullRequest, fn PullProgressFunc
 	})
 }
 
-// PushProgressFunc is a function that [Client.Push] invokes when progress is
-// made.
-// It's similar to other progress function types like [PullProgressFunc].
 type PushProgressFunc func(ProgressResponse) error
 
-// Push uploads a model to the model library; requires registering for ollama.ai
-// and adding a public key first. fn is called each time progress is made on
-// the request and can be used to display a progress bar, etc.
 func (c *Client) Push(ctx context.Context, req *PushRequest, fn PushProgressFunc) error {
 	return c.stream(ctx, http.MethodPost, "/api/push", req, func(bts []byte) error {
 		var resp ProgressResponse
@@ -268,18 +189,11 @@ func (c *Client) Push(ctx context.Context, req *PushRequest, fn PushProgressFunc
 	})
 }
 
-// CreateProgressFunc is a function that [Client.Create] invokes when progress
-// is made.
-// It's similar to other progress function types like [PullProgressFunc].
-type CreateProgressFunc func(ProgressResponse) error
+type CreateProgressFunc func(CreateProgress) error
 
-// Create creates a model from a [Modelfile]. fn is a progress function that
-// behaves similarly to other methods (see [Client.Pull]).
-//
-// [Modelfile]: https://github.com/ollama/ollama/blob/main/docs/modelfile.md
 func (c *Client) Create(ctx context.Context, req *CreateRequest, fn CreateProgressFunc) error {
 	return c.stream(ctx, http.MethodPost, "/api/create", req, func(bts []byte) error {
-		var resp ProgressResponse
+		var resp CreateProgress
 		if err := json.Unmarshal(bts, &resp); err != nil {
 			return err
 		}
@@ -288,7 +202,6 @@ func (c *Client) Create(ctx context.Context, req *CreateRequest, fn CreateProgre
 	})
 }
 
-// List lists models that are available locally.
 func (c *Client) List(ctx context.Context) (*ListResponse, error) {
 	var lr ListResponse
 	if err := c.do(ctx, http.MethodGet, "/api/tags", nil, &lr); err != nil {
@@ -297,83 +210,9 @@ func (c *Client) List(ctx context.Context) (*ListResponse, error) {
 	return &lr, nil
 }
 
-// List running models.
-func (c *Client) ListRunning(ctx context.Context) (*ProcessResponse, error) {
-	var lr ProcessResponse
-	if err := c.do(ctx, http.MethodGet, "/api/ps", nil, &lr); err != nil {
-		return nil, err
-	}
-	return &lr, nil
-}
-
-// Copy copies a model - creating a model with another name from an existing
-// model.
-func (c *Client) Copy(ctx context.Context, req *CopyRequest) error {
-	if err := c.do(ctx, http.MethodPost, "/api/copy", req, nil); err != nil {
-		return err
-	}
-	return nil
-}
-
-// Delete deletes a model and its data.
 func (c *Client) Delete(ctx context.Context, req *DeleteRequest) error {
 	if err := c.do(ctx, http.MethodDelete, "/api/delete", req, nil); err != nil {
 		return err
 	}
 	return nil
 }
-
-// Show obtains model information, including details, modelfile, license etc.
-func (c *Client) Show(ctx context.Context, req *ShowRequest) (*ShowResponse, error) {
-	var resp ShowResponse
-	if err := c.do(ctx, http.MethodPost, "/api/show", req, &resp); err != nil {
-		return nil, err
-	}
-	return &resp, nil
-}
-
-// Hearbeat checks if the server has started and is responsive; if yes, it
-// returns nil, otherwise an error.
-func (c *Client) Heartbeat(ctx context.Context) error {
-	if err := c.do(ctx, http.MethodHead, "/", nil, nil); err != nil {
-		return err
-	}
-	return nil
-}
-
-// Embed generates embeddings from a model.
-func (c *Client) Embed(ctx context.Context, req *EmbedRequest) (*EmbedResponse, error) {
-	var resp EmbedResponse
-	if err := c.do(ctx, http.MethodPost, "/api/embed", req, &resp); err != nil {
-		return nil, err
-	}
-	return &resp, nil
-}
-
-// Embeddings generates an embedding from a model.
-func (c *Client) Embeddings(ctx context.Context, req *EmbeddingRequest) (*EmbeddingResponse, error) {
-	var resp EmbeddingResponse
-	if err := c.do(ctx, http.MethodPost, "/api/embeddings", req, &resp); err != nil {
-		return nil, err
-	}
-	return &resp, nil
-}
-
-// CreateBlob creates a blob from a file on the server. digest is the
-// expected SHA256 digest of the file, and r represents the file.
-func (c *Client) CreateBlob(ctx context.Context, digest string, r io.Reader) error {
-	return c.do(ctx, http.MethodPost, fmt.Sprintf("/api/blobs/%s", digest), r, nil)
-}
-
-// Version returns the Ollama server version as a string.
-func (c *Client) Version(ctx context.Context) (string, error) {
-	var version struct {
-		Version string `json:"version"`
-	}
-
-	if err := c.do(ctx, http.MethodGet, "/api/version", nil, &version); err != nil {
-		return "", err
-	}
-
-	return version.Version, nil
-}

api/client_test.go

@@ -1,45 +0,0 @@
-package api
-
-import (
-	"testing"
-)
-
-func TestClientFromEnvironment(t *testing.T) {
-	type testCase struct {
-		value  string
-		expect string
-		err    error
-	}
-
-	testCases := map[string]*testCase{
-		"empty":                      {value: "", expect: "http://127.0.0.1:11434"},
-		"only address":               {value: "1.2.3.4", expect: "http://1.2.3.4:11434"},
-		"only port":                  {value: ":1234", expect: "http://:1234"},
-		"address and port":           {value: "1.2.3.4:1234", expect: "http://1.2.3.4:1234"},
-		"scheme http and address":    {value: "http://1.2.3.4", expect: "http://1.2.3.4:80"},
-		"scheme https and address":   {value: "https://1.2.3.4", expect: "https://1.2.3.4:443"},
-		"scheme, address, and port":  {value: "https://1.2.3.4:1234", expect: "https://1.2.3.4:1234"},
-		"hostname":                   {value: "example.com", expect: "http://example.com:11434"},
-		"hostname and port":          {value: "example.com:1234", expect: "http://example.com:1234"},
-		"scheme http and hostname":   {value: "http://example.com", expect: "http://example.com:80"},
-		"scheme https and hostname":  {value: "https://example.com", expect: "https://example.com:443"},
-		"scheme, hostname, and port": {value: "https://example.com:1234", expect: "https://example.com:1234"},
-		"trailing slash":             {value: "example.com/", expect: "http://example.com:11434"},
-		"trailing slash port":        {value: "example.com:1234/", expect: "http://example.com:1234"},
-	}
-
-	for k, v := range testCases {
-		t.Run(k, func(t *testing.T) {
-			t.Setenv("OLLAMA_HOST", v.value)
-
-			client, err := ClientFromEnvironment()
-			if err != v.err {
-				t.Fatalf("expected %s, got %s", v.err, err)
-			}
-
-			if client.base.String() != v.expect {
-				t.Fatalf("expected %s, got %s", v.expect, client.base.String())
-			}
-		})
-	}
-}

api/types.go

@@ -1,18 +1,12 @@
 package api
 
 import (
-	"encoding/json"
 	"fmt"
-	"log/slog"
-	"math"
 	"os"
-	"reflect"
-	"strconv"
-	"strings"
+	"runtime"
 	"time"
 )
 
-// StatusError is an error with and HTTP status code.
 type StatusError struct {
 	StatusCode   int
 	Status       string
@@ -33,701 +27,160 @@ func (e StatusError) Error() string {
 	}
 }
 
-// ImageData represents the raw binary data of an image file.
-type ImageData []byte
-
-// GenerateRequest describes a request sent by [Client.Generate]. While you
-// have to specify the Model and Prompt fields, all the other fields have
-// reasonable defaults for basic uses.
 type GenerateRequest struct {
-	// Model is the model name; it should be a name familiar to Ollama from
-	// the library at https://ollama.com/library
-	Model string `json:"model"`
+	Model   string `json:"model"`
+	Prompt  string `json:"prompt"`
+	Context []int  `json:"context,omitempty"`
 
-	// Prompt is the textual prompt to send to the model.
-	Prompt string `json:"prompt"`
+	Options `json:"options"`
+}
 
-	// Suffix is the text that comes after the inserted text.
-	Suffix string `json:"suffix"`
+type CreateRequest struct {
+	Name string `json:"name"`
+	Path string `json:"path"`
+}
 
-	// System overrides the model's default system message/prompt.
-	System string `json:"system"`
+type CreateProgress struct {
+	Status string `json:"status"`
+}
 
-	// Template overrides the model's default prompt template.
-	Template string `json:"template"`
+type DeleteRequest struct {
+	Name string `json:"name"`
+}
 
-	// Context is the context parameter returned from a previous call to
-	// Generate call. It can be used to keep a short conversational memory.
+type PullRequest struct {
+	Name     string `json:"name"`
+	Insecure bool   `json:"insecure,omitempty"`
+	Username string `json:"username"`
+	Password string `json:"password"`
+}
+
+type ProgressResponse struct {
+	Status    string `json:"status"`
+	Digest    string `json:"digest,omitempty"`
+	Total     int    `json:"total,omitempty"`
+	Completed int    `json:"completed,omitempty"`
+}
+
+type PushRequest struct {
+	Name     string `json:"name"`
+	Insecure bool   `json:"insecure,omitempty"`
+	Username string `json:"username"`
+	Password string `json:"password"`
+}
+
+type ListResponse struct {
+	Models []ListResponseModel `json:"models"`
+}
+
+type ListResponseModel struct {
+	Name       string    `json:"name"`
+	ModifiedAt time.Time `json:"modified_at"`
+	Size       int       `json:"size"`
+}
+
+type GenerateResponse struct {
+	Model     string    `json:"model"`
+	CreatedAt time.Time `json:"created_at"`
+	Response  string    `json:"response,omitempty"`
+
+	Done    bool  `json:"done"`
 	Context []int `json:"context,omitempty"`
 
-	// Stream specifies whether the response is streaming; it is true by default.
-	Stream *bool `json:"stream,omitempty"`
-
-	// Raw set to true means that no formatting will be applied to the prompt.
-	Raw bool `json:"raw,omitempty"`
-
-	// Format specifies the format to return a response in.
-	Format string `json:"format"`
-
-	// KeepAlive controls how long the model will stay loaded in memory following
-	// this request.
-	KeepAlive *Duration `json:"keep_alive,omitempty"`
-
-	// Images is an optional list of base64-encoded images accompanying this
-	// request, for multimodal models.
-	Images []ImageData `json:"images,omitempty"`
-
-	// Options lists model-specific options. For example, temperature can be
-	// set through this field, if the model supports it.
-	Options map[string]interface{} `json:"options"`
-}
-
-// ChatRequest describes a request sent by [Client.Chat].
-type ChatRequest struct {
-	// Model is the model name, as in [GenerateRequest].
-	Model string `json:"model"`
-
-	// Messages is the messages of the chat - can be used to keep a chat memory.
-	Messages []Message `json:"messages"`
-
-	// Stream enable streaming of returned response; true by default.
-	Stream *bool `json:"stream,omitempty"`
-
-	// Format is the format to return the response in (e.g. "json").
-	Format string `json:"format"`
-
-	// KeepAlive controls how long the model will stay loaded into memory
-	// followin the request.
-	KeepAlive *Duration `json:"keep_alive,omitempty"`
-
-	// Tools is an optional list of tools the model has access to.
-	Tools `json:"tools,omitempty"`
-
-	// Options lists model-specific options.
-	Options map[string]interface{} `json:"options"`
-}
-
-type Tools []Tool
-
-func (t Tools) String() string {
-	bts, _ := json.Marshal(t)
-	return string(bts)
-}
-
-func (t Tool) String() string {
-	bts, _ := json.Marshal(t)
-	return string(bts)
-}
-
-// Message is a single message in a chat sequence. The message contains the
-// role ("system", "user", or "assistant"), the content and an optional list
-// of images.
-type Message struct {
-	Role      string      `json:"role"`
-	Content   string      `json:"content"`
-	Images    []ImageData `json:"images,omitempty"`
-	ToolCalls []ToolCall  `json:"tool_calls,omitempty"`
-}
-
-func (m *Message) UnmarshalJSON(b []byte) error {
-	type Alias Message
-	var a Alias
-	if err := json.Unmarshal(b, &a); err != nil {
-		return err
-	}
-
-	*m = Message(a)
-	m.Role = strings.ToLower(m.Role)
-	return nil
-}
-
-type ToolCall struct {
-	Function ToolCallFunction `json:"function"`
-}
-
-type ToolCallFunction struct {
-	Name      string                    `json:"name"`
-	Arguments ToolCallFunctionArguments `json:"arguments"`
-}
-
-type ToolCallFunctionArguments map[string]any
-
-func (t *ToolCallFunctionArguments) String() string {
-	bts, _ := json.Marshal(t)
-	return string(bts)
-}
-
-type Tool struct {
-	Type     string       `json:"type"`
-	Function ToolFunction `json:"function"`
-}
-
-type ToolFunction struct {
-	Name        string `json:"name"`
-	Description string `json:"description"`
-	Parameters  struct {
-		Type       string   `json:"type"`
-		Required   []string `json:"required"`
-		Properties map[string]struct {
-			Type        string   `json:"type"`
-			Description string   `json:"description"`
-			Enum        []string `json:"enum,omitempty"`
-		} `json:"properties"`
-	} `json:"parameters"`
-}
-
-func (t *ToolFunction) String() string {
-	bts, _ := json.Marshal(t)
-	return string(bts)
-}
-
-// ChatResponse is the response returned by [Client.Chat]. Its fields are
-// similar to [GenerateResponse].
-type ChatResponse struct {
-	Model      string    `json:"model"`
-	CreatedAt  time.Time `json:"created_at"`
-	Message    Message   `json:"message"`
-	DoneReason string    `json:"done_reason,omitempty"`
-
-	Done bool `json:"done"`
-
-	Metrics
-}
-
-type Metrics struct {
 	TotalDuration      time.Duration `json:"total_duration,omitempty"`
-	LoadDuration       time.Duration `json:"load_duration,omitempty"`
 	PromptEvalCount    int           `json:"prompt_eval_count,omitempty"`
 	PromptEvalDuration time.Duration `json:"prompt_eval_duration,omitempty"`
 	EvalCount          int           `json:"eval_count,omitempty"`
 	EvalDuration       time.Duration `json:"eval_duration,omitempty"`
 }
 
-// Options specified in [GenerateRequest], if you add a new option here add it
-// to the API docs also.
+func (r *GenerateResponse) Summary() {
+	if r.TotalDuration > 0 {
+		fmt.Fprintf(os.Stderr, "total duration:       %v\n", r.TotalDuration)
+	}
+
+	if r.PromptEvalCount > 0 {
+		fmt.Fprintf(os.Stderr, "prompt eval count:    %d token(s)\n", r.PromptEvalCount)
+	}
+
+	if r.PromptEvalDuration > 0 {
+		fmt.Fprintf(os.Stderr, "prompt eval duration: %s\n", r.PromptEvalDuration)
+		fmt.Fprintf(os.Stderr, "prompt eval rate:     %.2f tokens/s\n", float64(r.PromptEvalCount)/r.PromptEvalDuration.Seconds())
+	}
+
+	if r.EvalCount > 0 {
+		fmt.Fprintf(os.Stderr, "eval count:           %d token(s)\n", r.EvalCount)
+	}
+
+	if r.EvalDuration > 0 {
+		fmt.Fprintf(os.Stderr, "eval duration:        %s\n", r.EvalDuration)
+		fmt.Fprintf(os.Stderr, "eval rate:            %.2f tokens/s\n", float64(r.EvalCount)/r.EvalDuration.Seconds())
+	}
+}
+
 type Options struct {
-	Runner
+	Seed int `json:"seed,omitempty"`
 
-	// Predict options used at runtime
-	NumKeep          int      `json:"num_keep,omitempty"`
-	Seed             int      `json:"seed,omitempty"`
-	NumPredict       int      `json:"num_predict,omitempty"`
-	TopK             int      `json:"top_k,omitempty"`
-	TopP             float32  `json:"top_p,omitempty"`
-	MinP             float32  `json:"min_p,omitempty"`
-	TFSZ             float32  `json:"tfs_z,omitempty"`
-	TypicalP         float32  `json:"typical_p,omitempty"`
-	RepeatLastN      int      `json:"repeat_last_n,omitempty"`
-	Temperature      float32  `json:"temperature,omitempty"`
-	RepeatPenalty    float32  `json:"repeat_penalty,omitempty"`
-	PresencePenalty  float32  `json:"presence_penalty,omitempty"`
-	FrequencyPenalty float32  `json:"frequency_penalty,omitempty"`
-	Mirostat         int      `json:"mirostat,omitempty"`
-	MirostatTau      float32  `json:"mirostat_tau,omitempty"`
-	MirostatEta      float32  `json:"mirostat_eta,omitempty"`
-	PenalizeNewline  bool     `json:"penalize_newline,omitempty"`
-	Stop             []string `json:"stop,omitempty"`
+	// Backend options
+	UseNUMA bool `json:"numa,omitempty"`
+
+	// Model options
+	NumCtx        int  `json:"num_ctx,omitempty"`
+	NumBatch      int  `json:"num_batch,omitempty"`
+	NumGPU        int  `json:"num_gpu,omitempty"`
+	MainGPU       int  `json:"main_gpu,omitempty"`
+	LowVRAM       bool `json:"low_vram,omitempty"`
+	F16KV         bool `json:"f16_kv,omitempty"`
+	LogitsAll     bool `json:"logits_all,omitempty"`
+	VocabOnly     bool `json:"vocab_only,omitempty"`
+	UseMMap       bool `json:"use_mmap,omitempty"`
+	UseMLock      bool `json:"use_mlock,omitempty"`
+	EmbeddingOnly bool `json:"embedding_only,omitempty"`
+
+	// Predict options
+	RepeatLastN      int     `json:"repeat_last_n,omitempty"`
+	RepeatPenalty    float32 `json:"repeat_penalty,omitempty"`
+	FrequencyPenalty float32 `json:"frequency_penalty,omitempty"`
+	PresencePenalty  float32 `json:"presence_penalty,omitempty"`
+	Temperature      float32 `json:"temperature,omitempty"`
+	TopK             int     `json:"top_k,omitempty"`
+	TopP             float32 `json:"top_p,omitempty"`
+	TFSZ             float32 `json:"tfs_z,omitempty"`
+	TypicalP         float32 `json:"typical_p,omitempty"`
+	Mirostat         int     `json:"mirostat,omitempty"`
+	MirostatTau      float32 `json:"mirostat_tau,omitempty"`
+	MirostatEta      float32 `json:"mirostat_eta,omitempty"`
+
+	NumThread int `json:"num_thread,omitempty"`
 }
 
-// Runner options which must be set when the model is loaded into memory
-type Runner struct {
-	UseNUMA   bool  `json:"numa,omitempty"`
-	NumCtx    int   `json:"num_ctx,omitempty"`
-	NumBatch  int   `json:"num_batch,omitempty"`
-	NumGPU    int   `json:"num_gpu,omitempty"`
-	MainGPU   int   `json:"main_gpu,omitempty"`
-	LowVRAM   bool  `json:"low_vram,omitempty"`
-	F16KV     bool  `json:"f16_kv,omitempty"`
-	LogitsAll bool  `json:"logits_all,omitempty"`
-	VocabOnly bool  `json:"vocab_only,omitempty"`
-	UseMMap   *bool `json:"use_mmap,omitempty"`
-	UseMLock  bool  `json:"use_mlock,omitempty"`
-	NumThread int   `json:"num_thread,omitempty"`
-}
-
-// EmbedRequest is the request passed to [Client.Embed].
-type EmbedRequest struct {
-	// Model is the model name.
-	Model string `json:"model"`
-
-	// Input is the input to embed.
-	Input any `json:"input"`
-
-	// KeepAlive controls how long the model will stay loaded in memory following
-	// this request.
-	KeepAlive *Duration `json:"keep_alive,omitempty"`
-
-	Truncate *bool `json:"truncate,omitempty"`
-
-	// Options lists model-specific options.
-	Options map[string]interface{} `json:"options"`
-}
-
-// EmbedResponse is the response from [Client.Embed].
-type EmbedResponse struct {
-	Model      string      `json:"model"`
-	Embeddings [][]float32 `json:"embeddings"`
-
-	TotalDuration   time.Duration `json:"total_duration,omitempty"`
-	LoadDuration    time.Duration `json:"load_duration,omitempty"`
-	PromptEvalCount int           `json:"prompt_eval_count,omitempty"`
-}
-
-// EmbeddingRequest is the request passed to [Client.Embeddings].
-type EmbeddingRequest struct {
-	// Model is the model name.
-	Model string `json:"model"`
-
-	// Prompt is the textual prompt to embed.
-	Prompt string `json:"prompt"`
-
-	// KeepAlive controls how long the model will stay loaded in memory following
-	// this request.
-	KeepAlive *Duration `json:"keep_alive,omitempty"`
-
-	// Options lists model-specific options.
-	Options map[string]interface{} `json:"options"`
-}
-
-// EmbeddingResponse is the response from [Client.Embeddings].
-type EmbeddingResponse struct {
-	Embedding []float64 `json:"embedding"`
-}
-
-// CreateRequest is the request passed to [Client.Create].
-type CreateRequest struct {
-	Model     string `json:"model"`
-	Path      string `json:"path"`
-	Modelfile string `json:"modelfile"`
-	Stream    *bool  `json:"stream,omitempty"`
-	Quantize  string `json:"quantize,omitempty"`
-
-	// Name is deprecated, see Model
-	Name string `json:"name"`
-
-	// Quantization is deprecated, see Quantize
-	Quantization string `json:"quantization,omitempty"`
-}
-
-// DeleteRequest is the request passed to [Client.Delete].
-type DeleteRequest struct {
-	Model string `json:"model"`
-
-	// Name is deprecated, see Model
-	Name string `json:"name"`
-}
-
-// ShowRequest is the request passed to [Client.Show].
-type ShowRequest struct {
-	Model  string `json:"model"`
-	System string `json:"system"`
-
-	// Template is deprecated
-	Template string `json:"template"`
-	Verbose  bool   `json:"verbose"`
-
-	Options map[string]interface{} `json:"options"`
-
-	// Name is deprecated, see Model
-	Name string `json:"name"`
-}
-
-// ShowResponse is the response returned from [Client.Show].
-type ShowResponse struct {
-	License       string         `json:"license,omitempty"`
-	Modelfile     string         `json:"modelfile,omitempty"`
-	Parameters    string         `json:"parameters,omitempty"`
-	Template      string         `json:"template,omitempty"`
-	System        string         `json:"system,omitempty"`
-	Details       ModelDetails   `json:"details,omitempty"`
-	Messages      []Message      `json:"messages,omitempty"`
-	ModelInfo     map[string]any `json:"model_info,omitempty"`
-	ProjectorInfo map[string]any `json:"projector_info,omitempty"`
-	ModifiedAt    time.Time      `json:"modified_at,omitempty"`
-}
-
-// CopyRequest is the request passed to [Client.Copy].
-type CopyRequest struct {
-	Source      string `json:"source"`
-	Destination string `json:"destination"`
-}
-
-// PullRequest is the request passed to [Client.Pull].
-type PullRequest struct {
-	Model    string `json:"model"`
-	Insecure bool   `json:"insecure,omitempty"`
-	Username string `json:"username"`
-	Password string `json:"password"`
-	Stream   *bool  `json:"stream,omitempty"`
-
-	// Name is deprecated, see Model
-	Name string `json:"name"`
-}
-
-// ProgressResponse is the response passed to progress functions like
-// [PullProgressFunc] and [PushProgressFunc].
-type ProgressResponse struct {
-	Status    string `json:"status"`
-	Digest    string `json:"digest,omitempty"`
-	Total     int64  `json:"total,omitempty"`
-	Completed int64  `json:"completed,omitempty"`
-}
-
-// PushRequest is the request passed to [Client.Push].
-type PushRequest struct {
-	Model    string `json:"model"`
-	Insecure bool   `json:"insecure,omitempty"`
-	Username string `json:"username"`
-	Password string `json:"password"`
-	Stream   *bool  `json:"stream,omitempty"`
-
-	// Name is deprecated, see Model
-	Name string `json:"name"`
-}
-
-// ListResponse is the response from [Client.List].
-type ListResponse struct {
-	Models []ListModelResponse `json:"models"`
-}
-
-// ProcessResponse is the response from [Client.Process].
-type ProcessResponse struct {
-	Models []ProcessModelResponse `json:"models"`
-}
-
-// ListModelResponse is a single model description in [ListResponse].
-type ListModelResponse struct {
-	Name       string       `json:"name"`
-	Model      string       `json:"model"`
-	ModifiedAt time.Time    `json:"modified_at"`
-	Size       int64        `json:"size"`
-	Digest     string       `json:"digest"`
-	Details    ModelDetails `json:"details,omitempty"`
-}
-
-// ProcessModelResponse is a single model description in [ProcessResponse].
-type ProcessModelResponse struct {
-	Name      string       `json:"name"`
-	Model     string       `json:"model"`
-	Size      int64        `json:"size"`
-	Digest    string       `json:"digest"`
-	Details   ModelDetails `json:"details,omitempty"`
-	ExpiresAt time.Time    `json:"expires_at"`
-	SizeVRAM  int64        `json:"size_vram"`
-}
-
-type RetrieveModelResponse struct {
-	Id      string `json:"id"`
-	Object  string `json:"object"`
-	Created int64  `json:"created"`
-	OwnedBy string `json:"owned_by"`
-}
-
-type TokenResponse struct {
-	Token string `json:"token"`
-}
-
-// GenerateResponse is the response passed into [GenerateResponseFunc].
-type GenerateResponse struct {
-	// Model is the model name that generated the response.
-	Model string `json:"model"`
-
-	// CreatedAt is the timestamp of the response.
-	CreatedAt time.Time `json:"created_at"`
-
-	// Response is the textual response itself.
-	Response string `json:"response"`
-
-	// Done specifies if the response is complete.
-	Done bool `json:"done"`
-
-	// DoneReason is the reason the model stopped generating text.
-	DoneReason string `json:"done_reason,omitempty"`
-
-	// Context is an encoding of the conversation used in this response; this
-	// can be sent in the next request to keep a conversational memory.
-	Context []int `json:"context,omitempty"`
-
-	Metrics
-}
-
-// ModelDetails provides details about a model.
-type ModelDetails struct {
-	ParentModel       string   `json:"parent_model"`
-	Format            string   `json:"format"`
-	Family            string   `json:"family"`
-	Families          []string `json:"families"`
-	ParameterSize     string   `json:"parameter_size"`
-	QuantizationLevel string   `json:"quantization_level"`
-}
-
-func (m *Metrics) Summary() {
-	if m.TotalDuration > 0 {
-		fmt.Fprintf(os.Stderr, "total duration:       %v\n", m.TotalDuration)
-	}
-
-	if m.LoadDuration > 0 {
-		fmt.Fprintf(os.Stderr, "load duration:        %v\n", m.LoadDuration)
-	}
-
-	if m.PromptEvalCount > 0 {
-		fmt.Fprintf(os.Stderr, "prompt eval count:    %d token(s)\n", m.PromptEvalCount)
-	}
-
-	if m.PromptEvalDuration > 0 {
-		fmt.Fprintf(os.Stderr, "prompt eval duration: %s\n", m.PromptEvalDuration)
-		fmt.Fprintf(os.Stderr, "prompt eval rate:     %.2f tokens/s\n", float64(m.PromptEvalCount)/m.PromptEvalDuration.Seconds())
-	}
-
-	if m.EvalCount > 0 {
-		fmt.Fprintf(os.Stderr, "eval count:           %d token(s)\n", m.EvalCount)
-	}
-
-	if m.EvalDuration > 0 {
-		fmt.Fprintf(os.Stderr, "eval duration:        %s\n", m.EvalDuration)
-		fmt.Fprintf(os.Stderr, "eval rate:            %.2f tokens/s\n", float64(m.EvalCount)/m.EvalDuration.Seconds())
-	}
-}
-
-func (opts *Options) FromMap(m map[string]interface{}) error {
-	valueOpts := reflect.ValueOf(opts).Elem() // names of the fields in the options struct
-	typeOpts := reflect.TypeOf(opts).Elem()   // types of the fields in the options struct
-
-	// build map of json struct tags to their types
-	jsonOpts := make(map[string]reflect.StructField)
-	for _, field := range reflect.VisibleFields(typeOpts) {
-		jsonTag := strings.Split(field.Tag.Get("json"), ",")[0]
-		if jsonTag != "" {
-			jsonOpts[jsonTag] = field
-		}
-	}
-
-	for key, val := range m {
-		opt, ok := jsonOpts[key]
-		if !ok {
-			slog.Warn("invalid option provided", "option", opt.Name)
-			continue
-		}
-
-		field := valueOpts.FieldByName(opt.Name)
-		if field.IsValid() && field.CanSet() {
-			if val == nil {
-				continue
-			}
-
-			switch field.Kind() {
-			case reflect.Int:
-				switch t := val.(type) {
-				case int64:
-					field.SetInt(t)
-				case float64:
-					// when JSON unmarshals numbers, it uses float64, not int
-					field.SetInt(int64(t))
-				default:
-					return fmt.Errorf("option %q must be of type integer", key)
-				}
-			case reflect.Bool:
-				val, ok := val.(bool)
-				if !ok {
-					return fmt.Errorf("option %q must be of type boolean", key)
-				}
-				field.SetBool(val)
-			case reflect.Float32:
-				// JSON unmarshals to float64
-				val, ok := val.(float64)
-				if !ok {
-					return fmt.Errorf("option %q must be of type float32", key)
-				}
-				field.SetFloat(val)
-			case reflect.String:
-				val, ok := val.(string)
-				if !ok {
-					return fmt.Errorf("option %q must be of type string", key)
-				}
-				field.SetString(val)
-			case reflect.Slice:
-				// JSON unmarshals to []interface{}, not []string
-				val, ok := val.([]interface{})
-				if !ok {
-					return fmt.Errorf("option %q must be of type array", key)
-				}
-				// convert []interface{} to []string
-				slice := make([]string, len(val))
-				for i, item := range val {
-					str, ok := item.(string)
-					if !ok {
-						return fmt.Errorf("option %q must be of an array of strings", key)
-					}
-					slice[i] = str
-				}
-				field.Set(reflect.ValueOf(slice))
-			case reflect.Pointer:
-				var b bool
-				if field.Type() == reflect.TypeOf(&b) {
-					val, ok := val.(bool)
-					if !ok {
-						return fmt.Errorf("option %q must be of type boolean", key)
-					}
-					field.Set(reflect.ValueOf(&val))
-				} else {
-					return fmt.Errorf("unknown type loading config params: %v %v", field.Kind(), field.Type())
-				}
-			default:
-				return fmt.Errorf("unknown type loading config params: %v", field.Kind())
-			}
-		}
-	}
-
-	return nil
-}
-
-// DefaultOptions is the default set of options for [GenerateRequest]; these
-// values are used unless the user specifies other values explicitly.
 func DefaultOptions() Options {
 	return Options{
-		// options set on request to runner
-		NumPredict: -1,
+		Seed: -1,
 
-		// set a minimal num_keep to avoid issues on context shifts
-		NumKeep:          4,
+		UseNUMA: false,
+
+		NumCtx:   2048,
+		NumBatch: 512,
+		NumGPU:   1,
+		LowVRAM:  false,
+		F16KV:    true,
+		UseMMap:  true,
+		UseMLock: false,
+
+		RepeatLastN:      512,
+		RepeatPenalty:    1.1,
+		FrequencyPenalty: 0.0,
+		PresencePenalty:  0.0,
 		Temperature:      0.8,
 		TopK:             40,
 		TopP:             0.9,
 		TFSZ:             1.0,
 		TypicalP:         1.0,
-		RepeatLastN:      64,
-		RepeatPenalty:    1.1,
-		PresencePenalty:  0.0,
-		FrequencyPenalty: 0.0,
 		Mirostat:         0,
 		MirostatTau:      5.0,
 		MirostatEta:      0.1,
-		PenalizeNewline:  true,
-		Seed:             -1,
 
-		Runner: Runner{
-			// options set when the model is loaded
-			NumCtx:    2048,
-			NumBatch:  512,
-			NumGPU:    -1, // -1 here indicates that NumGPU should be set dynamically
-			NumThread: 0,  // let the runtime decide
-			LowVRAM:   false,
-			F16KV:     true,
-			UseMLock:  false,
-			UseMMap:   nil,
-			UseNUMA:   false,
-		},
+		NumThread: runtime.NumCPU(),
 	}
 }
-
-type Duration struct {
-	time.Duration
-}
-
-func (d Duration) MarshalJSON() ([]byte, error) {
-	if d.Duration < 0 {
-		return []byte("-1"), nil
-	}
-	return []byte("\"" + d.Duration.String() + "\""), nil
-}
-
-func (d *Duration) UnmarshalJSON(b []byte) (err error) {
-	var v any
-	if err := json.Unmarshal(b, &v); err != nil {
-		return err
-	}
-
-	d.Duration = 5 * time.Minute
-
-	switch t := v.(type) {
-	case float64:
-		if t < 0 {
-			d.Duration = time.Duration(math.MaxInt64)
-		} else {
-			d.Duration = time.Duration(int(t) * int(time.Second))
-		}
-	case string:
-		d.Duration, err = time.ParseDuration(t)
-		if err != nil {
-			return err
-		}
-		if d.Duration < 0 {
-			d.Duration = time.Duration(math.MaxInt64)
-		}
-	default:
-		return fmt.Errorf("Unsupported type: '%s'", reflect.TypeOf(v))
-	}
-
-	return nil
-}
-
-// FormatParams converts specified parameter options to their correct types
-func FormatParams(params map[string][]string) (map[string]interface{}, error) {
-	opts := Options{}
-	valueOpts := reflect.ValueOf(&opts).Elem() // names of the fields in the options struct
-	typeOpts := reflect.TypeOf(opts)           // types of the fields in the options struct
-
-	// build map of json struct tags to their types
-	jsonOpts := make(map[string]reflect.StructField)
-	for _, field := range reflect.VisibleFields(typeOpts) {
-		jsonTag := strings.Split(field.Tag.Get("json"), ",")[0]
-		if jsonTag != "" {
-			jsonOpts[jsonTag] = field
-		}
-	}
-
-	out := make(map[string]interface{})
-	// iterate params and set values based on json struct tags
-	for key, vals := range params {
-		if opt, ok := jsonOpts[key]; !ok {
-			return nil, fmt.Errorf("unknown parameter '%s'", key)
-		} else {
-			field := valueOpts.FieldByName(opt.Name)
-			if field.IsValid() && field.CanSet() {
-				switch field.Kind() {
-				case reflect.Float32:
-					floatVal, err := strconv.ParseFloat(vals[0], 32)
-					if err != nil {
-						return nil, fmt.Errorf("invalid float value %s", vals)
-					}
-
-					out[key] = float32(floatVal)
-				case reflect.Int:
-					intVal, err := strconv.ParseInt(vals[0], 10, 64)
-					if err != nil {
-						return nil, fmt.Errorf("invalid int value %s", vals)
-					}
-
-					out[key] = intVal
-				case reflect.Bool:
-					boolVal, err := strconv.ParseBool(vals[0])
-					if err != nil {
-						return nil, fmt.Errorf("invalid bool value %s", vals)
-					}
-
-					out[key] = boolVal
-				case reflect.String:
-					out[key] = vals[0]
-				case reflect.Slice:
-					// TODO: only string slices are supported right now
-					out[key] = vals
-				case reflect.Pointer:
-					var b bool
-					if field.Type() == reflect.TypeOf(&b) {
-						boolVal, err := strconv.ParseBool(vals[0])
-						if err != nil {
-							return nil, fmt.Errorf("invalid bool value %s", vals)
-						}
-						out[key] = &boolVal
-					} else {
-						return nil, fmt.Errorf("unknown type %s for %s", field.Kind(), key)
-					}
-				default:
-					return nil, fmt.Errorf("unknown type %s for %s", field.Kind(), key)
-				}
-			}
-		}
-	}
-
-	return out, nil
-}

api/types_test.go

@@ -1,233 +0,0 @@
-package api
-
-import (
-	"encoding/json"
-	"fmt"
-	"math"
-	"testing"
-	"time"
-
-	"github.com/stretchr/testify/assert"
-	"github.com/stretchr/testify/require"
-)
-
-func TestKeepAliveParsingFromJSON(t *testing.T) {
-	tests := []struct {
-		name string
-		req  string
-		exp  *Duration
-	}{
-		{
-			name: "Positive Integer",
-			req:  `{ "keep_alive": 42 }`,
-			exp:  &Duration{42 * time.Second},
-		},
-		{
-			name: "Positive Float",
-			req:  `{ "keep_alive": 42.5 }`,
-			exp:  &Duration{42 * time.Second},
-		},
-		{
-			name: "Positive Integer String",
-			req:  `{ "keep_alive": "42m" }`,
-			exp:  &Duration{42 * time.Minute},
-		},
-		{
-			name: "Negative Integer",
-			req:  `{ "keep_alive": -1 }`,
-			exp:  &Duration{math.MaxInt64},
-		},
-		{
-			name: "Negative Float",
-			req:  `{ "keep_alive": -3.14 }`,
-			exp:  &Duration{math.MaxInt64},
-		},
-		{
-			name: "Negative Integer String",
-			req:  `{ "keep_alive": "-1m" }`,
-			exp:  &Duration{math.MaxInt64},
-		},
-	}
-
-	for _, test := range tests {
-		t.Run(test.name, func(t *testing.T) {
-			var dec ChatRequest
-			err := json.Unmarshal([]byte(test.req), &dec)
-			require.NoError(t, err)
-
-			assert.Equal(t, test.exp, dec.KeepAlive)
-		})
-	}
-}
-
-func TestDurationMarshalUnmarshal(t *testing.T) {
-	tests := []struct {
-		name     string
-		input    time.Duration
-		expected time.Duration
-	}{
-		{
-			"negative duration",
-			time.Duration(-1),
-			time.Duration(math.MaxInt64),
-		},
-		{
-			"positive duration",
-			42 * time.Second,
-			42 * time.Second,
-		},
-		{
-			"another positive duration",
-			42 * time.Minute,
-			42 * time.Minute,
-		},
-		{
-			"zero duration",
-			time.Duration(0),
-			time.Duration(0),
-		},
-		{
-			"max duration",
-			time.Duration(math.MaxInt64),
-			time.Duration(math.MaxInt64),
-		},
-	}
-
-	for _, test := range tests {
-		t.Run(test.name, func(t *testing.T) {
-			b, err := json.Marshal(Duration{test.input})
-			require.NoError(t, err)
-
-			var d Duration
-			err = json.Unmarshal(b, &d)
-			require.NoError(t, err)
-
-			assert.Equal(t, test.expected, d.Duration, "input %v, marshalled %v, got %v", test.input, string(b), d.Duration)
-		})
-	}
-}
-
-func TestUseMmapParsingFromJSON(t *testing.T) {
-	tr := true
-	fa := false
-	tests := []struct {
-		name string
-		req  string
-		exp  *bool
-	}{
-		{
-			name: "Undefined",
-			req:  `{ }`,
-			exp:  nil,
-		},
-		{
-			name: "True",
-			req:  `{ "use_mmap": true }`,
-			exp:  &tr,
-		},
-		{
-			name: "False",
-			req:  `{ "use_mmap": false }`,
-			exp:  &fa,
-		},
-	}
-
-	for _, test := range tests {
-		t.Run(test.name, func(t *testing.T) {
-			var oMap map[string]interface{}
-			err := json.Unmarshal([]byte(test.req), &oMap)
-			require.NoError(t, err)
-			opts := DefaultOptions()
-			err = opts.FromMap(oMap)
-			require.NoError(t, err)
-			assert.Equal(t, test.exp, opts.UseMMap)
-		})
-	}
-}
-
-func TestUseMmapFormatParams(t *testing.T) {
-	tr := true
-	fa := false
-	tests := []struct {
-		name string
-		req  map[string][]string
-		exp  *bool
-		err  error
-	}{
-		{
-			name: "True",
-			req: map[string][]string{
-				"use_mmap": {"true"},
-			},
-			exp: &tr,
-			err: nil,
-		},
-		{
-			name: "False",
-			req: map[string][]string{
-				"use_mmap": {"false"},
-			},
-			exp: &fa,
-			err: nil,
-		},
-		{
-			name: "Numeric True",
-			req: map[string][]string{
-				"use_mmap": {"1"},
-			},
-			exp: &tr,
-			err: nil,
-		},
-		{
-			name: "Numeric False",
-			req: map[string][]string{
-				"use_mmap": {"0"},
-			},
-			exp: &fa,
-			err: nil,
-		},
-		{
-			name: "invalid string",
-			req: map[string][]string{
-				"use_mmap": {"foo"},
-			},
-			exp: nil,
-			err: fmt.Errorf("invalid bool value [foo]"),
-		},
-	}
-
-	for _, test := range tests {
-		t.Run(test.name, func(t *testing.T) {
-			resp, err := FormatParams(test.req)
-			require.Equal(t, test.err, err)
-			respVal, ok := resp["use_mmap"]
-			if test.exp != nil {
-				assert.True(t, ok, "resp: %v", resp)
-				assert.Equal(t, *test.exp, *respVal.(*bool))
-			}
-		})
-	}
-}
-
-func TestMessage_UnmarshalJSON(t *testing.T) {
-	tests := []struct {
-		input    string
-		expected string
-	}{
-		{`{"role": "USER", "content": "Hello!"}`, "user"},
-		{`{"role": "System", "content": "Initialization complete."}`, "system"},
-		{`{"role": "assistant", "content": "How can I help you?"}`, "assistant"},
-		{`{"role": "TOOl", "content": "Access granted."}`, "tool"},
-	}
-
-	for _, test := range tests {
-		var msg Message
-		if err := json.Unmarshal([]byte(test.input), &msg); err != nil {
-			t.Errorf("Unexpected error: %v", err)
-		}
-
-		if msg.Role != test.expected {
-			t.Errorf("role not lowercased: got %v, expected %v", msg.Role, test.expected)
-		}
-	}
-}

app/.gitignore

@@ -1 +1,92 @@
-ollama.syso
+# Logs
+logs
+*.log
+npm-debug.log*
+yarn-debug.log*
+yarn-error.log*
+lerna-debug.log*
+
+# Diagnostic reports (https://nodejs.org/api/report.html)
+report.[0-9]*.[0-9]*.[0-9]*.[0-9]*.json
+
+# Runtime data
+pids
+*.pid
+*.seed
+*.pid.lock
+.DS_Store
+
+# Directory for instrumented libs generated by jscoverage/JSCover
+lib-cov
+
+# Coverage directory used by tools like istanbul
+coverage
+*.lcov
+
+# nyc test coverage
+.nyc_output
+
+# node-waf configuration
+.lock-wscript
+
+# Compiled binary addons (https://nodejs.org/api/addons.html)
+build/Release
+
+# Dependency directories
+node_modules/
+jspm_packages/
+
+# TypeScript v1 declaration files
+typings/
+
+# TypeScript cache
+*.tsbuildinfo
+
+# Optional npm cache directory
+.npm
+
+# Optional eslint cache
+.eslintcache
+
+# Optional REPL history
+.node_repl_history
+
+# Output of 'npm pack'
+*.tgz
+
+# Yarn Integrity file
+.yarn-integrity
+
+# dotenv environment variables file
+.env
+.env.test
+
+# parcel-bundler cache (https://parceljs.org/)
+.cache
+
+# next.js build output
+.next
+
+# nuxt.js build output
+.nuxt
+
+# vuepress build output
+.vuepress/dist
+
+# Serverless directories
+.serverless/
+
+# FuseBox cache
+.fusebox/
+
+# DynamoDB Local files
+.dynamodb/
+
+# Webpack
+.webpack/
+
+# Vite
+.vite/
+
+# Electron-Forge
+out/

app/README.md

@@ -1,22 +1,27 @@
-# Ollama App
+# Desktop
 
-## Linux
+_Note: the Ollama desktop app is a work in progress and is not ready yet for general use._
 
-TODO
+This app builds upon Ollama to provide a desktop experience for running models.
 
-## MacOS
+## Developing
 
-TODO
-
-## Windows
-
-If you want to build the installer, youll need to install
-- https://jrsoftware.org/isinfo.php
-
-
-In the top directory of this repo, run the following powershell script
-to build the ollama CLI, ollama app, and ollama installer.
+First, build the `ollama` binary:
 
 ```
-powershell -ExecutionPolicy Bypass -File .\scripts\build_windows.ps1
+make -C ..
 ```
+
+Then run the desktop app with `npm start`:
+
+```
+npm install
+npm start
+```
+
+## Coming soon
+
+- Browse the latest available models on Hugging Face and other sources
+- Keep track of previous conversations with models
+- Switch quickly between models
+- Connect to remote Ollama servers to run models

app/assets/assets.go

@@ -1,17 +0,0 @@
-package assets
-
-import (
-	"embed"
-	"io/fs"
-)
-
-//go:embed *.ico
-var icons embed.FS
-
-func ListIcons() ([]string, error) {
-	return fs.Glob(icons, "*")
-}
-
-func GetIcon(filename string) ([]byte, error) {
-	return icons.ReadFile(filename)
-}

app/forge.config.ts

@@ -18,15 +18,12 @@ const config: ForgeConfig = {
     asar: true,
     icon: './assets/icon.icns',
     extraResource: [
-      '../dist/ollama',
-      path.join(__dirname, './assets/iconTemplate.png'),
-      path.join(__dirname, './assets/iconTemplate@2x.png'),
-      path.join(__dirname, './assets/iconUpdateTemplate.png'),
-      path.join(__dirname, './assets/iconUpdateTemplate@2x.png'),
-      path.join(__dirname, './assets/iconDarkTemplate.png'),
-      path.join(__dirname, './assets/iconDarkTemplate@2x.png'),
-      path.join(__dirname, './assets/iconDarkUpdateTemplate.png'),
-      path.join(__dirname, './assets/iconDarkUpdateTemplate@2x.png'),
+      '../ollama',
+      path.join(__dirname, './assets/ollama_icon_16x16Template.png'),
+      path.join(__dirname, './assets/ollama_icon_16x16Template@2x.png'),
+      path.join(__dirname, './assets/ollama_outline_icon_16x16Template.png'),
+      path.join(__dirname, './assets/ollama_outline_icon_16x16Template@2x.png'),
+      ...(process.platform === 'darwin' ? ['../llama/ggml-metal.metal'] : []),
     ],
     ...(process.env.SIGN
       ? {
@@ -41,12 +38,19 @@ const config: ForgeConfig = {
           },
         }
       : {}),
-    osxUniversal: {
-      x64ArchFiles: '**/ollama',
-    },
   },
   rebuildConfig: {},
   makers: [new MakerSquirrel({}), new MakerZIP({}, ['darwin'])],
+  publishers: [
+    new PublisherGithub({
+      repository: {
+        name: 'ollama',
+        owner: 'jmorganca',
+      },
+      draft: false,
+      prerelease: true,
+    }),
+  ],
   hooks: {
     readPackageJson: async (_, packageJson) => {
       return { ...packageJson, version: process.env.VERSION || packageJson.version }

app/lifecycle/getstarted_nonwindows.go

@@ -1,9 +0,0 @@
-//go:build !windows
-
-package lifecycle
-
-import "fmt"
-
-func GetStarted() error {
-	return fmt.Errorf("GetStarted not implemented")
-}

app/lifecycle/getstarted_windows.go

@@ -1,44 +0,0 @@
-package lifecycle
-
-import (
-	"fmt"
-	"log/slog"
-	"os"
-	"os/exec"
-	"path/filepath"
-	"syscall"
-)
-
-func GetStarted() error {
-	const CREATE_NEW_CONSOLE = 0x00000010
-	var err error
-	bannerScript := filepath.Join(AppDir, "ollama_welcome.ps1")
-	args := []string{
-		// TODO once we're signed, the execution policy bypass should be removed
-		"powershell", "-noexit", "-ExecutionPolicy", "Bypass", "-nologo", "-file", bannerScript,
-	}
-	args[0], err = exec.LookPath(args[0])
-	if err != nil {
-		return err
-	}
-
-	// Make sure the script actually exists
-	_, err = os.Stat(bannerScript)
-	if err != nil {
-		return fmt.Errorf("getting started banner script error %s", err)
-	}
-
-	slog.Info(fmt.Sprintf("opening getting started terminal with %v", args))
-	attrs := &os.ProcAttr{
-		Files: []*os.File{os.Stdin, os.Stdout, os.Stderr},
-		Sys:   &syscall.SysProcAttr{CreationFlags: CREATE_NEW_CONSOLE, HideWindow: false},
-	}
-	proc, err := os.StartProcess(args[0], args, attrs)
-
-	if err != nil {
-		return fmt.Errorf("unable to start getting started shell %w", err)
-	}
-
-	slog.Debug(fmt.Sprintf("getting started terminal PID: %d", proc.Pid))
-	return proc.Release()
-}

app/lifecycle/lifecycle.go

@@ -1,92 +0,0 @@
-package lifecycle
-
-import (
-	"context"
-	"fmt"
-	"log"
-	"log/slog"
-	"os"
-	"os/signal"
-	"syscall"
-
-	"github.com/ollama/ollama/app/store"
-	"github.com/ollama/ollama/app/tray"
-)
-
-func Run() {
-	InitLogging()
-
-	ctx, cancel := context.WithCancel(context.Background())
-	var done chan int
-
-	t, err := tray.NewTray()
-	if err != nil {
-		log.Fatalf("Failed to start: %s", err)
-	}
-	callbacks := t.GetCallbacks()
-
-	signals := make(chan os.Signal, 1)
-	signal.Notify(signals, syscall.SIGINT, syscall.SIGTERM)
-
-	go func() {
-		slog.Debug("starting callback loop")
-		for {
-			select {
-			case <-callbacks.Quit:
-				slog.Debug("quit called")
-				t.Quit()
-			case <-signals:
-				slog.Debug("shutting down due to signal")
-				t.Quit()
-			case <-callbacks.Update:
-				err := DoUpgrade(cancel, done)
-				if err != nil {
-					slog.Warn(fmt.Sprintf("upgrade attempt failed: %s", err))
-				}
-			case <-callbacks.ShowLogs:
-				ShowLogs()
-			case <-callbacks.DoFirstUse:
-				err := GetStarted()
-				if err != nil {
-					slog.Warn(fmt.Sprintf("Failed to launch getting started shell: %s", err))
-				}
-			}
-		}
-	}()
-
-	// Are we first use?
-	if !store.GetFirstTimeRun() {
-		slog.Debug("First time run")
-		err = t.DisplayFirstUseNotification()
-		if err != nil {
-			slog.Debug(fmt.Sprintf("XXX failed to display first use notification %v", err))
-		}
-		store.SetFirstTimeRun(true)
-	} else {
-		slog.Debug("Not first time, skipping first run notification")
-	}
-
-	if IsServerRunning(ctx) {
-		slog.Info("Detected another instance of ollama running, exiting")
-		os.Exit(1)
-	} else {
-		done, err = SpawnServer(ctx, CLIName)
-		if err != nil {
-			// TODO - should we retry in a backoff loop?
-			// TODO - should we pop up a warning and maybe add a menu item to view application logs?
-			slog.Error(fmt.Sprintf("Failed to spawn ollama server %s", err))
-			done = make(chan int, 1)
-			done <- 1
-		}
-	}
-
-	StartBackgroundUpdaterChecker(ctx, t.UpdateAvailable)
-
-	t.Run()
-	cancel()
-	slog.Info("Waiting for ollama server to shutdown...")
-	if done != nil {
-		<-done
-	}
-	slog.Info("Ollama app exiting")
-}

app/lifecycle/logging.go

@@ -1,80 +0,0 @@
-package lifecycle
-
-import (
-	"fmt"
-	"log/slog"
-	"os"
-	"path/filepath"
-	"strconv"
-	"strings"
-
-	"github.com/ollama/ollama/envconfig"
-)
-
-func InitLogging() {
-	level := slog.LevelInfo
-
-	if envconfig.Debug() {
-		level = slog.LevelDebug
-	}
-
-	var logFile *os.File
-	var err error
-	// Detect if we're a GUI app on windows, and if not, send logs to console
-	if os.Stderr.Fd() != 0 {
-		// Console app detected
-		logFile = os.Stderr
-		// TODO - write one-line to the app.log file saying we're running in console mode to help avoid confusion
-	} else {
-		rotateLogs(AppLogFile)
-		logFile, err = os.OpenFile(AppLogFile, os.O_APPEND|os.O_WRONLY|os.O_CREATE, 0755)
-		if err != nil {
-			slog.Error(fmt.Sprintf("failed to create server log %v", err))
-			return
-		}
-	}
-	handler := slog.NewTextHandler(logFile, &slog.HandlerOptions{
-		Level:     level,
-		AddSource: true,
-		ReplaceAttr: func(_ []string, attr slog.Attr) slog.Attr {
-			if attr.Key == slog.SourceKey {
-				source := attr.Value.Any().(*slog.Source)
-				source.File = filepath.Base(source.File)
-			}
-			return attr
-		},
-	})
-
-	slog.SetDefault(slog.New(handler))
-
-	slog.Info("ollama app started")
-}
-
-func rotateLogs(logFile string) {
-	if _, err := os.Stat(logFile); os.IsNotExist(err) {
-		return
-	}
-	index := strings.LastIndex(logFile, ".")
-	pre := logFile[:index]
-	post := "." + logFile[index+1:]
-	for i := LogRotationCount; i > 0; i-- {
-		older := pre + "-" + strconv.Itoa(i) + post
-		newer := pre + "-" + strconv.Itoa(i-1) + post
-		if i == 1 {
-			newer = pre + post
-		}
-		if _, err := os.Stat(newer); err == nil {
-			if _, err := os.Stat(older); err == nil {
-				err := os.Remove(older)
-				if err != nil {
-					slog.Warn("Failed to remove older log", "older", older, "error", err)
-					continue
-				}
-			}
-			err := os.Rename(newer, older)
-			if err != nil {
-				slog.Warn("Failed to rotate log", "older", older, "newer", newer, "error", err)
-			}
-		}
-	}
-}

app/lifecycle/logging_nonwindows.go

@@ -1,9 +0,0 @@
-//go:build !windows
-
-package lifecycle
-
-import "log/slog"
-
-func ShowLogs() {
-	slog.Warn("ShowLogs not yet implemented")
-}

app/lifecycle/logging_test.go

@@ -1,44 +0,0 @@
-package lifecycle
-
-import (
-	"os"
-	"path/filepath"
-	"strconv"
-	"testing"
-
-	"github.com/stretchr/testify/assert"
-	"github.com/stretchr/testify/require"
-)
-
-func TestRotateLogs(t *testing.T) {
-	logDir := t.TempDir()
-	logFile := filepath.Join(logDir, "testlog.log")
-
-	// No log exists
-	rotateLogs(logFile)
-
-	require.NoError(t, os.WriteFile(logFile, []byte("1"), 0644))
-	assert.FileExists(t, logFile)
-	// First rotation
-	rotateLogs(logFile)
-	assert.FileExists(t, filepath.Join(logDir, "testlog-1.log"))
-	assert.NoFileExists(t, filepath.Join(logDir, "testlog-2.log"))
-	assert.NoFileExists(t, logFile)
-
-	// Should be a no-op without a new log
-	rotateLogs(logFile)
-	assert.FileExists(t, filepath.Join(logDir, "testlog-1.log"))
-	assert.NoFileExists(t, filepath.Join(logDir, "testlog-2.log"))
-	assert.NoFileExists(t, logFile)
-
-	for i := 2; i <= LogRotationCount+1; i++ {
-		require.NoError(t, os.WriteFile(logFile, []byte(strconv.Itoa(i)), 0644))
-		assert.FileExists(t, logFile)
-		rotateLogs(logFile)
-		assert.NoFileExists(t, logFile)
-		for j := 1; j < i; j++ {
-			assert.FileExists(t, filepath.Join(logDir, "testlog-"+strconv.Itoa(j)+".log"))
-		}
-		assert.NoFileExists(t, filepath.Join(logDir, "testlog-"+strconv.Itoa(i+1)+".log"))
-	}
-}

app/lifecycle/logging_windows.go

@@ -1,19 +0,0 @@
-package lifecycle
-
-import (
-	"fmt"
-	"log/slog"
-	"os/exec"
-	"syscall"
-)
-
-func ShowLogs() {
-	cmd_path := "c:\\Windows\\system32\\cmd.exe"
-	slog.Debug(fmt.Sprintf("viewing logs with start %s", AppDataDir))
-	cmd := exec.Command(cmd_path, "/c", "start", AppDataDir)
-	cmd.SysProcAttr = &syscall.SysProcAttr{HideWindow: false, CreationFlags: 0x08000000}
-	err := cmd.Start()
-	if err != nil {
-		slog.Error(fmt.Sprintf("Failed to open log dir: %s", err))
-	}
-}

app/lifecycle/paths.go

@@ -1,79 +0,0 @@
-package lifecycle
-
-import (
-	"errors"
-	"fmt"
-	"log/slog"
-	"os"
-	"path/filepath"
-	"runtime"
-	"strings"
-)
-
-var (
-	AppName    = "ollama app"
-	CLIName    = "ollama"
-	AppDir     = "/opt/Ollama"
-	AppDataDir = "/opt/Ollama"
-	// TODO - should there be a distinct log dir?
-	UpdateStageDir   = "/tmp"
-	AppLogFile       = "/tmp/ollama_app.log"
-	ServerLogFile    = "/tmp/ollama.log"
-	UpgradeLogFile   = "/tmp/ollama_update.log"
-	Installer        = "OllamaSetup.exe"
-	LogRotationCount = 5
-)
-
-func init() {
-	if runtime.GOOS == "windows" {
-		AppName += ".exe"
-		CLIName += ".exe"
-		// Logs, configs, downloads go to LOCALAPPDATA
-		localAppData := os.Getenv("LOCALAPPDATA")
-		AppDataDir = filepath.Join(localAppData, "Ollama")
-		UpdateStageDir = filepath.Join(AppDataDir, "updates")
-		AppLogFile = filepath.Join(AppDataDir, "app.log")
-		ServerLogFile = filepath.Join(AppDataDir, "server.log")
-		UpgradeLogFile = filepath.Join(AppDataDir, "upgrade.log")
-
-		// Executables are stored in APPDATA
-		AppDir = filepath.Join(localAppData, "Programs", "Ollama")
-
-		// Make sure we have PATH set correctly for any spawned children
-		paths := strings.Split(os.Getenv("PATH"), ";")
-		// Start with whatever we find in the PATH/LD_LIBRARY_PATH
-		found := false
-		for _, path := range paths {
-			d, err := filepath.Abs(path)
-			if err != nil {
-				continue
-			}
-			if strings.EqualFold(AppDir, d) {
-				found = true
-			}
-		}
-		if !found {
-			paths = append(paths, AppDir)
-
-			pathVal := strings.Join(paths, ";")
-			slog.Debug("setting PATH=" + pathVal)
-			err := os.Setenv("PATH", pathVal)
-			if err != nil {
-				slog.Error(fmt.Sprintf("failed to update PATH: %s", err))
-			}
-		}
-
-		// Make sure our logging dir exists
-		_, err := os.Stat(AppDataDir)
-		if errors.Is(err, os.ErrNotExist) {
-			if err := os.MkdirAll(AppDataDir, 0o755); err != nil {
-				slog.Error(fmt.Sprintf("create ollama dir %s: %v", AppDataDir, err))
-			}
-		}
-	} else if runtime.GOOS == "darwin" {
-		// TODO
-		AppName += ".app"
-		// } else if runtime.GOOS == "linux" {
-		// TODO
-	}
-}

app/lifecycle/server.go

@@ -1,180 +0,0 @@
-package lifecycle
-
-import (
-	"context"
-	"errors"
-	"fmt"
-	"io"
-	"log/slog"
-	"os"
-	"os/exec"
-	"path/filepath"
-	"time"
-
-	"github.com/ollama/ollama/api"
-)
-
-func getCLIFullPath(command string) string {
-	var cmdPath string
-	appExe, err := os.Executable()
-	if err == nil {
-		cmdPath = filepath.Join(filepath.Dir(appExe), command)
-		_, err := os.Stat(cmdPath)
-		if err == nil {
-			return cmdPath
-		}
-	}
-	cmdPath, err = exec.LookPath(command)
-	if err == nil {
-		_, err := os.Stat(cmdPath)
-		if err == nil {
-			return cmdPath
-		}
-	}
-	pwd, err := os.Getwd()
-	if err == nil {
-		cmdPath = filepath.Join(pwd, command)
-		_, err = os.Stat(cmdPath)
-		if err == nil {
-			return cmdPath
-		}
-	}
-
-	return command
-}
-
-func start(ctx context.Context, command string) (*exec.Cmd, error) {
-	cmd := getCmd(ctx, getCLIFullPath(command))
-	stdout, err := cmd.StdoutPipe()
-	if err != nil {
-		return nil, fmt.Errorf("failed to spawn server stdout pipe: %w", err)
-	}
-	stderr, err := cmd.StderrPipe()
-	if err != nil {
-		return nil, fmt.Errorf("failed to spawn server stderr pipe: %w", err)
-	}
-
-	rotateLogs(ServerLogFile)
-	logFile, err := os.OpenFile(ServerLogFile, os.O_APPEND|os.O_WRONLY|os.O_CREATE, 0755)
-	if err != nil {
-		return nil, fmt.Errorf("failed to create server log: %w", err)
-	}
-
-	logDir := filepath.Dir(ServerLogFile)
-	_, err = os.Stat(logDir)
-	if err != nil {
-		if !errors.Is(err, os.ErrNotExist) {
-			return nil, fmt.Errorf("stat ollama server log dir %s: %v", logDir, err)
-		}
-
-		if err := os.MkdirAll(logDir, 0o755); err != nil {
-			return nil, fmt.Errorf("create ollama server log dir %s: %v", logDir, err)
-		}
-	}
-
-	go func() {
-		defer logFile.Close()
-		io.Copy(logFile, stdout) //nolint:errcheck
-	}()
-	go func() {
-		defer logFile.Close()
-		io.Copy(logFile, stderr) //nolint:errcheck
-	}()
-
-	// Re-wire context done behavior to attempt a graceful shutdown of the server
-	cmd.Cancel = func() error {
-		if cmd.Process != nil {
-			err := terminate(cmd)
-			if err != nil {
-				slog.Warn("error trying to gracefully terminate server", "err", err)
-				return cmd.Process.Kill()
-			}
-
-			tick := time.NewTicker(10 * time.Millisecond)
-			defer tick.Stop()
-
-			for {
-				select {
-				case <-tick.C:
-					exited, err := isProcessExited(cmd.Process.Pid)
-					if err != nil {
-						return err
-					}
-
-					if exited {
-						return nil
-					}
-				case <-time.After(5 * time.Second):
-					slog.Warn("graceful server shutdown timeout, killing", "pid", cmd.Process.Pid)
-					return cmd.Process.Kill()
-				}
-			}
-		}
-		return nil
-	}
-
-	// run the command and wait for it to finish
-	if err := cmd.Start(); err != nil {
-		return nil, fmt.Errorf("failed to start server %w", err)
-	}
-	if cmd.Process != nil {
-		slog.Info(fmt.Sprintf("started ollama server with pid %d", cmd.Process.Pid))
-	}
-	slog.Info(fmt.Sprintf("ollama server logs %s", ServerLogFile))
-
-	return cmd, nil
-}
-
-func SpawnServer(ctx context.Context, command string) (chan int, error) {
-	done := make(chan int)
-
-	go func() {
-		// Keep the server running unless we're shuttind down the app
-		crashCount := 0
-		for {
-			slog.Info("starting server...")
-			cmd, err := start(ctx, command)
-			if err != nil {
-				crashCount++
-				slog.Error(fmt.Sprintf("failed to start server %s", err))
-				time.Sleep(500 * time.Millisecond * time.Duration(crashCount))
-				continue
-			}
-
-			cmd.Wait() //nolint:errcheck
-			var code int
-			if cmd.ProcessState != nil {
-				code = cmd.ProcessState.ExitCode()
-			}
-
-			select {
-			case <-ctx.Done():
-				slog.Info(fmt.Sprintf("server shutdown with exit code %d", code))
-				done <- code
-				return
-			default:
-				crashCount++
-				slog.Warn(fmt.Sprintf("server crash %d - exit code %d - respawning", crashCount, code))
-				time.Sleep(500 * time.Millisecond * time.Duration(crashCount))
-				break
-			}
-		}
-	}()
-
-	return done, nil
-}
-
-func IsServerRunning(ctx context.Context) bool {
-	client, err := api.ClientFromEnvironment()
-	if err != nil {
-		slog.Info("unable to connect to server")
-		return false
-	}
-	err = client.Heartbeat(ctx)
-	if err != nil {
-		slog.Debug(fmt.Sprintf("heartbeat from server: %s", err))
-		slog.Info("unable to connect to server")
-		return false
-	}
-	return true
-}

app/lifecycle/server_unix.go

@@ -1,38 +0,0 @@
-//go:build !windows
-
-package lifecycle
-
-import (
-	"context"
-	"errors"
-	"fmt"
-	"os"
-	"os/exec"
-	"syscall"
-)
-
-func getCmd(ctx context.Context, cmd string) *exec.Cmd {
-	return exec.CommandContext(ctx, cmd, "serve")
-}
-
-func terminate(cmd *exec.Cmd) error {
-	return cmd.Process.Signal(os.Interrupt)
-}
-
-func isProcessExited(pid int) (bool, error) {
-	proc, err := os.FindProcess(pid)
-	if err != nil {
-		return false, fmt.Errorf("failed to find process: %v", err)
-	}
-
-	err = proc.Signal(syscall.Signal(0))
-	if err != nil {
-		if errors.Is(err, os.ErrProcessDone) || errors.Is(err, syscall.ESRCH) {
-			return true, nil
-		}
-
-		return false, fmt.Errorf("error signaling process: %v", err)
-	}
-
-	return false, nil
-}

app/lifecycle/server_windows.go

@@ -1,91 +0,0 @@
-package lifecycle
-
-import (
-	"context"
-	"fmt"
-	"os/exec"
-	"syscall"
-
-	"golang.org/x/sys/windows"
-)
-
-func getCmd(ctx context.Context, exePath string) *exec.Cmd {
-	cmd := exec.CommandContext(ctx, exePath, "serve")
-	cmd.SysProcAttr = &syscall.SysProcAttr{
-		HideWindow:    true,
-		CreationFlags: windows.CREATE_NEW_PROCESS_GROUP,
-	}
-
-	return cmd
-}
-
-func terminate(cmd *exec.Cmd) error {
-	dll, err := windows.LoadDLL("kernel32.dll")
-	if err != nil {
-		return err
-	}
-	//nolint:errcheck
-	defer dll.Release()
-
-	pid := cmd.Process.Pid
-
-	f, err := dll.FindProc("AttachConsole")
-	if err != nil {
-		return err
-	}
-
-	r1, _, err := f.Call(uintptr(pid))
-	if r1 == 0 && err != syscall.ERROR_ACCESS_DENIED {
-		return err
-	}
-
-	f, err = dll.FindProc("SetConsoleCtrlHandler")
-	if err != nil {
-		return err
-	}
-
-	r1, _, err = f.Call(0, 1)
-	if r1 == 0 {
-		return err
-	}
-
-	f, err = dll.FindProc("GenerateConsoleCtrlEvent")
-	if err != nil {
-		return err
-	}
-
-	r1, _, err = f.Call(windows.CTRL_BREAK_EVENT, uintptr(pid))
-	if r1 == 0 {
-		return err
-	}
-
-	r1, _, err = f.Call(windows.CTRL_C_EVENT, uintptr(pid))
-	if r1 == 0 {
-		return err
-	}
-
-	return nil
-}
-
-const STILL_ACTIVE = 259
-
-func isProcessExited(pid int) (bool, error) {
-	hProcess, err := windows.OpenProcess(windows.PROCESS_QUERY_INFORMATION, false, uint32(pid))
-	if err != nil {
-		return false, fmt.Errorf("failed to open process: %v", err)
-	}
-	//nolint:errcheck
-	defer windows.CloseHandle(hProcess)
-
-	var exitCode uint32
-	err = windows.GetExitCodeProcess(hProcess, &exitCode)
-	if err != nil {
-		return false, fmt.Errorf("failed to get exit code: %v", err)
-	}
-
-	if exitCode == STILL_ACTIVE {
-		return false, nil
-	}
-
-	return true, nil
-}

app/lifecycle/updater.go

@@ -1,228 +0,0 @@
-package lifecycle
-
-import (
-	"context"
-	"crypto/rand"
-	"encoding/json"
-	"errors"
-	"fmt"
-	"io"
-	"log/slog"
-	"mime"
-	"net/http"
-	"net/url"
-	"os"
-	"path"
-	"path/filepath"
-	"runtime"
-	"strings"
-	"time"
-
-	"github.com/ollama/ollama/auth"
-	"github.com/ollama/ollama/version"
-)
-
-var (
-	UpdateCheckURLBase  = "https://ollama.com/api/update"
-	UpdateDownloaded    = false
-	UpdateCheckInterval = 60 * 60 * time.Second
-)
-
-// TODO - maybe move up to the API package?
-type UpdateResponse struct {
-	UpdateURL     string `json:"url"`
-	UpdateVersion string `json:"version"`
-}
-
-func IsNewReleaseAvailable(ctx context.Context) (bool, UpdateResponse) {
-	var updateResp UpdateResponse
-
-	requestURL, err := url.Parse(UpdateCheckURLBase)
-	if err != nil {
-		return false, updateResp
-	}
-
-	query := requestURL.Query()
-	query.Add("os", runtime.GOOS)
-	query.Add("arch", runtime.GOARCH)
-	query.Add("version", version.Version)
-	query.Add("ts", fmt.Sprintf("%d", time.Now().Unix()))
-
-	nonce, err := auth.NewNonce(rand.Reader, 16)
-	if err != nil {
-		return false, updateResp
-	}
-
-	query.Add("nonce", nonce)
-	requestURL.RawQuery = query.Encode()
-
-	data := []byte(fmt.Sprintf("%s,%s", http.MethodGet, requestURL.RequestURI()))
-	signature, err := auth.Sign(ctx, data)
-	if err != nil {
-		return false, updateResp
-	}
-
-	req, err := http.NewRequestWithContext(ctx, http.MethodGet, requestURL.String(), nil)
-	if err != nil {
-		slog.Warn(fmt.Sprintf("failed to check for update: %s", err))
-		return false, updateResp
-	}
-	req.Header.Set("Authorization", signature)
-	req.Header.Set("User-Agent", fmt.Sprintf("ollama/%s (%s %s) Go/%s", version.Version, runtime.GOARCH, runtime.GOOS, runtime.Version()))
-
-	slog.Debug("checking for available update", "requestURL", requestURL)
-	resp, err := http.DefaultClient.Do(req)
-	if err != nil {
-		slog.Warn(fmt.Sprintf("failed to check for update: %s", err))
-		return false, updateResp
-	}
-	defer resp.Body.Close()
-
-	if resp.StatusCode == http.StatusNoContent {
-		slog.Debug("check update response 204 (current version is up to date)")
-		return false, updateResp
-	}
-	body, err := io.ReadAll(resp.Body)
-	if err != nil {
-		slog.Warn(fmt.Sprintf("failed to read body response: %s", err))
-	}
-
-	if resp.StatusCode != http.StatusOK {
-		slog.Info(fmt.Sprintf("check update error %d - %.96s", resp.StatusCode, string(body)))
-		return false, updateResp
-	}
-	err = json.Unmarshal(body, &updateResp)
-	if err != nil {
-		slog.Warn(fmt.Sprintf("malformed response checking for update: %s", err))
-		return false, updateResp
-	}
-	// Extract the version string from the URL in the github release artifact path
-	updateResp.UpdateVersion = path.Base(path.Dir(updateResp.UpdateURL))
-
-	slog.Info("New update available at " + updateResp.UpdateURL)
-	return true, updateResp
-}
-
-func DownloadNewRelease(ctx context.Context, updateResp UpdateResponse) error {
-	// Do a head first to check etag info
-	req, err := http.NewRequestWithContext(ctx, http.MethodHead, updateResp.UpdateURL, nil)
-	if err != nil {
-		return err
-	}
-
-	resp, err := http.DefaultClient.Do(req)
-	if err != nil {
-		return fmt.Errorf("error checking update: %w", err)
-	}
-	if resp.StatusCode != http.StatusOK {
-		return fmt.Errorf("unexpected status attempting to download update %d", resp.StatusCode)
-	}
-	resp.Body.Close()
-	etag := strings.Trim(resp.Header.Get("etag"), "\"")
-	if etag == "" {
-		slog.Debug("no etag detected, falling back to filename based dedup")
-		etag = "_"
-	}
-	filename := Installer
-	_, params, err := mime.ParseMediaType(resp.Header.Get("content-disposition"))
-	if err == nil {
-		filename = params["filename"]
-	}
-
-	stageFilename := filepath.Join(UpdateStageDir, etag, filename)
-
-	// Check to see if we already have it downloaded
-	_, err = os.Stat(stageFilename)
-	if err == nil {
-		slog.Info("update already downloaded")
-		return nil
-	}
-
-	cleanupOldDownloads()
-
-	req.Method = http.MethodGet
-	resp, err = http.DefaultClient.Do(req)
-	if err != nil {
-		return fmt.Errorf("error checking update: %w", err)
-	}
-	defer resp.Body.Close()
-	etag = strings.Trim(resp.Header.Get("etag"), "\"")
-	if etag == "" {
-		slog.Debug("no etag detected, falling back to filename based dedup") // TODO probably can get rid of this redundant log
-		etag = "_"
-	}
-
-	stageFilename = filepath.Join(UpdateStageDir, etag, filename)
-
-	_, err = os.Stat(filepath.Dir(stageFilename))
-	if errors.Is(err, os.ErrNotExist) {
-		if err := os.MkdirAll(filepath.Dir(stageFilename), 0o755); err != nil {
-			return fmt.Errorf("create ollama dir %s: %v", filepath.Dir(stageFilename), err)
-		}
-	}
-
-	payload, err := io.ReadAll(resp.Body)
-	if err != nil {
-		return fmt.Errorf("failed to read body response: %w", err)
-	}
-	fp, err := os.OpenFile(stageFilename, os.O_WRONLY|os.O_CREATE|os.O_TRUNC, 0o755)
-	if err != nil {
-		return fmt.Errorf("write payload %s: %w", stageFilename, err)
-	}
-	defer fp.Close()
-	if n, err := fp.Write(payload); err != nil || n != len(payload) {
-		return fmt.Errorf("write payload %s: %d vs %d -- %w", stageFilename, n, len(payload), err)
-	}
-	slog.Info("new update downloaded " + stageFilename)
-
-	UpdateDownloaded = true
-	return nil
-}
-
-func cleanupOldDownloads() {
-	files, err := os.ReadDir(UpdateStageDir)
-	if err != nil && errors.Is(err, os.ErrNotExist) {
-		// Expected behavior on first run
-		return
-	} else if err != nil {
-		slog.Warn(fmt.Sprintf("failed to list stage dir: %s", err))
-		return
-	}
-	for _, file := range files {
-		fullname := filepath.Join(UpdateStageDir, file.Name())
-		slog.Debug("cleaning up old download: " + fullname)
-		err = os.RemoveAll(fullname)
-		if err != nil {
-			slog.Warn(fmt.Sprintf("failed to cleanup stale update download %s", err))
-		}
-	}
-}
-
-func StartBackgroundUpdaterChecker(ctx context.Context, cb func(string) error) {
-	go func() {
-		// Don't blast an update message immediately after startup
-		// time.Sleep(30 * time.Second)
-		time.Sleep(3 * time.Second)
-
-		for {
-			available, resp := IsNewReleaseAvailable(ctx)
-			if available {
-				err := DownloadNewRelease(ctx, resp)
-				if err != nil {
-					slog.Error(fmt.Sprintf("failed to download new release: %s", err))
-				}
-				err = cb(resp.UpdateVersion)
-				if err != nil {
-					slog.Warn(fmt.Sprintf("failed to register update available with tray: %s", err))
-				}
-			}
-			select {
-			case <-ctx.Done():
-				slog.Debug("stopping background update checker")
-				return
-			default:
-				time.Sleep(UpdateCheckInterval)
-			}
-		}
-	}()
-}

app/lifecycle/updater_nonwindows.go

@@ -1,12 +0,0 @@
-//go:build !windows
-
-package lifecycle
-
-import (
-	"context"
-	"fmt"
-)
-
-func DoUpgrade(cancel context.CancelFunc, done chan int) error {
-	return fmt.Errorf("DoUpgrade not yet implemented")
-}

app/lifecycle/updater_windows.go

@@ -1,77 +0,0 @@
-package lifecycle
-
-import (
-	"context"
-	"fmt"
-	"log/slog"
-	"os"
-	"os/exec"
-	"path/filepath"
-)
-
-func DoUpgrade(cancel context.CancelFunc, done chan int) error {
-	files, err := filepath.Glob(filepath.Join(UpdateStageDir, "*", "*.exe")) // TODO generalize for multiplatform
-	if err != nil {
-		return fmt.Errorf("failed to lookup downloads: %s", err)
-	}
-	if len(files) == 0 {
-		return fmt.Errorf("no update downloads found")
-	} else if len(files) > 1 {
-		// Shouldn't happen
-		slog.Warn(fmt.Sprintf("multiple downloads found, using first one %v", files))
-	}
-	installerExe := files[0]
-
-	slog.Info("starting upgrade with " + installerExe)
-	slog.Info("upgrade log file " + UpgradeLogFile)
-
-	// When running in debug mode, we'll be "verbose" and let the installer pop up and prompt
-	installArgs := []string{
-		"/CLOSEAPPLICATIONS",                    // Quit the tray app if it's still running
-		"/LOG=" + filepath.Base(UpgradeLogFile), // Only relative seems reliable, so set pwd
-		"/FORCECLOSEAPPLICATIONS",               // Force close the tray app - might be needed
-	}
-	// make the upgrade as quiet as possible (no GUI, no prompts)
-	installArgs = append(installArgs,
-		"/SP", // Skip the "This will install... Do you wish to continue" prompt
-		"/SUPPRESSMSGBOXES",
-		"/SILENT",
-		"/VERYSILENT",
-	)
-
-	// Safeguard in case we have requests in flight that need to drain...
-	slog.Info("Waiting for server to shutdown")
-	cancel()
-	if done != nil {
-		<-done
-	} else {
-		// Shouldn't happen
-		slog.Warn("done chan was nil, not actually waiting")
-	}
-
-	slog.Debug(fmt.Sprintf("starting installer: %s %v", installerExe, installArgs))
-	os.Chdir(filepath.Dir(UpgradeLogFile)) //nolint:errcheck
-	cmd := exec.Command(installerExe, installArgs...)
-
-	if err := cmd.Start(); err != nil {
-		return fmt.Errorf("unable to start ollama app %w", err)
-	}
-
-	if cmd.Process != nil {
-		err = cmd.Process.Release()
-		if err != nil {
-			slog.Error(fmt.Sprintf("failed to release server process: %s", err))
-		}
-	} else {
-		// TODO - some details about why it didn't start, or is this a pedantic error case?
-		return fmt.Errorf("installer process did not start")
-	}
-
-	// TODO should we linger for a moment and check to make sure it's actually running by checking the pid?
-
-	slog.Info("Installer started in background, exiting")
-
-	os.Exit(0)
-	// Not reached
-	return nil
-}

app/main.go

@@ -1,12 +0,0 @@
-package main
-
-// Compile with the following to get rid of the cmd pop up on windows
-// go build -ldflags="-H windowsgui" .
-
-import (
-	"github.com/ollama/ollama/app/lifecycle"
-)
-
-func main() {
-	lifecycle.Run()
-}

app/ollama.iss

@@ -1,165 +0,0 @@
-; Inno Setup Installer for Ollama
-;
-; To build the installer use the build script invoked from the top of the source tree
-; 
-; powershell -ExecutionPolicy Bypass -File .\scripts\build_windows.ps
-
-
-#define MyAppName "Ollama"
-#if GetEnv("PKG_VERSION") != ""
-  #define MyAppVersion GetEnv("PKG_VERSION")
-#else
-  #define MyAppVersion "0.0.0"
-#endif
-#define MyAppPublisher "Ollama"
-#define MyAppURL "https://ollama.com/"
-#define MyAppExeName "ollama app.exe"
-#define MyIcon ".\assets\app.ico"
-
-[Setup]
-; NOTE: The value of AppId uniquely identifies this application. Do not use the same AppId value in installers for other applications.
-; (To generate a new GUID, click Tools | Generate GUID inside the IDE.)
-AppId={{44E83376-CE68-45EB-8FC1-393500EB558C}
-AppName={#MyAppName}
-AppVersion={#MyAppVersion}
-VersionInfoVersion={#MyAppVersion}
-;AppVerName={#MyAppName} {#MyAppVersion}
-AppPublisher={#MyAppPublisher}
-AppPublisherURL={#MyAppURL}
-AppSupportURL={#MyAppURL}
-AppUpdatesURL={#MyAppURL}
-ArchitecturesAllowed=x64 arm64
-ArchitecturesInstallIn64BitMode=x64 arm64
-DefaultDirName={localappdata}\Programs\{#MyAppName}
-DefaultGroupName={#MyAppName}
-DisableProgramGroupPage=yes
-PrivilegesRequired=lowest
-OutputBaseFilename="OllamaSetup"
-SetupIconFile={#MyIcon}
-UninstallDisplayIcon={uninstallexe}
-Compression=lzma2
-SolidCompression=no
-WizardStyle=modern
-ChangesEnvironment=yes
-OutputDir=..\dist\
-
-; Disable logging once everything's battle tested
-; Filename will be %TEMP%\Setup Log*.txt
-SetupLogging=yes
-CloseApplications=yes
-RestartApplications=no
-
-; https://jrsoftware.org/ishelp/index.php?topic=setup_wizardimagefile
-WizardSmallImageFile=.\assets\setup.bmp
-
-; TODO verifty actual min windows version...
-; OG Win 10
-MinVersion=10.0.10240
-
-; First release that supports WinRT UI Composition for win32 apps
-; MinVersion=10.0.17134
-; First release with XAML Islands - possible UI path forward
-; MinVersion=10.0.18362
-
-; quiet...
-DisableDirPage=yes
-DisableFinishedPage=yes
-DisableReadyMemo=yes
-DisableReadyPage=yes
-DisableStartupPrompt=yes
-DisableWelcomePage=yes
-
-; TODO - percentage can't be set less than 100, so how to make it shorter?
-; WizardSizePercent=100,80
-
-#if GetEnv("KEY_CONTAINER")
-SignTool=MySignTool
-SignedUninstaller=yes
-#endif
-
-SetupMutex=OllamaSetupMutex
-
-[Languages]
-Name: "english"; MessagesFile: "compiler:Default.isl"
-
-[LangOptions]
-DialogFontSize=12
-
-[Files]
-Source: ".\app.exe"; DestDir: "{app}"; DestName: "{#MyAppExeName}" ; Flags: ignoreversion 64bit
-Source: "..\ollama.exe"; DestDir: "{app}"; Flags: ignoreversion 64bit
-Source: "..\dist\windows-{#ARCH}\ollama_runners\*"; DestDir: "{app}\ollama_runners"; Flags: ignoreversion 64bit recursesubdirs
-Source: "..\dist\ollama_welcome.ps1"; DestDir: "{app}"; Flags: ignoreversion
-Source: ".\assets\app.ico"; DestDir: "{app}"; Flags: ignoreversion
-#if DirExists("..\dist\windows-amd64\cuda")
-  Source: "..\dist\windows-amd64\cuda\*"; DestDir: "{app}\cuda\"; Flags: ignoreversion recursesubdirs
-#endif
-#if DirExists("..\dist\windows-amd64\oneapi")
-  Source: "..\dist\windows-amd64\oneapi\*"; DestDir: "{app}\oneapi\"; Flags: ignoreversion recursesubdirs
-#endif
-#if DirExists("..\dist\windows-amd64\rocm")
-  Source: "..\dist\windows-amd64\rocm\*"; DestDir: "{app}\rocm\"; Flags: ignoreversion recursesubdirs
-#endif
-
-
-[Icons]
-Name: "{group}\{#MyAppName}"; Filename: "{app}\{#MyAppExeName}"; IconFilename: "{app}\app.ico"
-Name: "{userstartup}\{#MyAppName}"; Filename: "{app}\{#MyAppExeName}"; IconFilename: "{app}\app.ico"
-Name: "{userprograms}\{#MyAppName}"; Filename: "{app}\{#MyAppExeName}"; IconFilename: "{app}\app.ico"
-
-[Run]
-Filename: "{cmd}"; Parameters: "/C set PATH={app};%PATH% & ""{app}\{#MyAppExeName}"""; Flags: postinstall nowait runhidden
-
-[UninstallRun]
-; Filename: "{cmd}"; Parameters: "/C ""taskkill /im ''{#MyAppExeName}'' /f /t"; Flags: runhidden
-; Filename: "{cmd}"; Parameters: "/C ""taskkill /im ollama.exe /f /t"; Flags: runhidden
-Filename: "taskkill"; Parameters: "/im ""{#MyAppExeName}"" /f /t"; Flags: runhidden
-Filename: "taskkill"; Parameters: "/im ""ollama.exe"" /f /t"; Flags: runhidden
-; HACK!  need to give the server and app enough time to exit
-; TODO - convert this to a Pascal code script so it waits until they're no longer running, then completes
-Filename: "{cmd}"; Parameters: "/c timeout 5"; Flags: runhidden
-
-[UninstallDelete]
-Type: filesandordirs; Name: "{%TEMP}\ollama*"
-Type: filesandordirs; Name: "{%LOCALAPPDATA}\Ollama"
-Type: filesandordirs; Name: "{%LOCALAPPDATA}\Programs\Ollama"
-Type: filesandordirs; Name: "{%USERPROFILE}\.ollama\models"
-Type: filesandordirs; Name: "{%USERPROFILE}\.ollama\history"
-; NOTE: if the user has a custom OLLAMA_MODELS it will be preserved
-
-[InstallDelete]
-Type: filesandordirs; Name: "{%TEMP}\ollama*"
-Type: filesandordirs; Name: "{%LOCALAPPDATA}\Programs\Ollama"
-
-[Messages]
-WizardReady=Ollama Windows Preview
-ReadyLabel1=%nLet's get you up and running with your own large language models.
-SetupAppRunningError=Another Ollama installer is running.%n%nPlease cancel or finish the other installer, then click OK to continue with this install, or Cancel to exit.
-
-
-;FinishedHeadingLabel=Run your first model
-;FinishedLabel=%nRun this command in a PowerShell or cmd terminal.%n%n%n    ollama run llama3.1
-;ClickFinish=%n
-
-[Registry]
-Root: HKCU; Subkey: "Environment"; \
-    ValueType: expandsz; ValueName: "Path"; ValueData: "{olddata};{app}"; \
-    Check: NeedsAddPath('{app}')
-
-[Code]
-
-function NeedsAddPath(Param: string): boolean;
-var
-  OrigPath: string;
-begin
-  if not RegQueryStringValue(HKEY_CURRENT_USER,
-    'Environment',
-    'Path', OrigPath)
-  then begin
-    Result := True;
-    exit;
-  end;
-  { look for the path with leading and trailing semicolon }
-  { Pos() returns 0 if not found }
-  Result := Pos(';' + ExpandConstant(Param) + ';', ';' + OrigPath + ';') = 0;
-end;

app/ollama.rc

@@ -1,29 +0,0 @@
-#include <winver.h>
-
-VS_VERSION_INFO VERSIONINFO
- FILEFLAGSMASK 0x3fL
-#ifdef _DEBUG
- FILEFLAGS 0x1L
-#else
- FILEFLAGS 0x0L
-#endif
- FILEOS 0x40004L
- FILETYPE 0x1L
- FILESUBTYPE 0x0L
-BEGIN
-    BLOCK "StringFileInfo"
-    BEGIN
-        BLOCK "040904b0"
-        BEGIN
-            VALUE "FileDescription", "Ollama"
-            VALUE "InternalName", "Ollama"
-            VALUE "OriginalFilename", "ollama app.exe"
-            VALUE "ProductName", "Ollama"
-        END
-    END
-
-    BLOCK "VarFileInfo"
-    BEGIN
-        VALUE "Translation", 0x409, 1200
-    END
-END

app/ollama_welcome.ps1

@@ -1,8 +0,0 @@
-# TODO - consider ANSI colors and maybe ASCII art...
-write-host ""
-write-host "Welcome to Ollama!"
-write-host ""
-write-host "Run your first model:"
-write-host ""
-write-host "`tollama run llama3.1"
-write-host ""

app/package.json

@@ -6,14 +6,12 @@
   "main": ".webpack/main",
   "scripts": {
     "start": "electron-forge start",
-    "package": "electron-forge package --arch universal",
-    "package:sign": "SIGN=1 electron-forge package --arch universal",
-    "make": "electron-forge make --arch universal",
-    "make:sign": "SIGN=1 electron-forge make --arch universal",
+    "package": "electron-forge package",
+    "package:sign": "SIGN=1 electron-forge package",
+    "make": "electron-forge make",
+    "make:sign": "SIGN=1 electron-forge make",
     "publish": "SIGN=1 electron-forge publish",
-    "lint": "eslint --ext .ts,.tsx .",
-    "format": "prettier --check . --ignore-path .gitignore",
-    "format:fix": "prettier --write . --ignore-path .gitignore"
+    "lint": "eslint --ext .ts,.tsx ."
   },
   "keywords": [],
   "author": {
@@ -32,7 +30,6 @@
     "@electron-forge/plugin-auto-unpack-natives": "^6.2.1",
     "@electron-forge/plugin-webpack": "^6.2.1",
     "@electron-forge/publisher-github": "^6.2.1",
-    "@electron/universal": "^1.4.1",
     "@svgr/webpack": "^8.0.1",
     "@types/chmodr": "^1.0.0",
     "@types/node": "^20.4.0",
@@ -46,7 +43,7 @@
     "chmodr": "^1.2.0",
     "copy-webpack-plugin": "^11.0.0",
     "css-loader": "^6.8.1",
-    "electron": "25.9.2",
+    "electron": "25.2.0",
     "eslint": "^8.43.0",
     "eslint-plugin-import": "^2.27.5",
     "fork-ts-checker-webpack-plugin": "^7.3.0",

app/src/app.tsx

@@ -2,7 +2,7 @@ import { useState } from 'react'
 import copy from 'copy-to-clipboard'
 import { CheckIcon, DocumentDuplicateIcon } from '@heroicons/react/24/outline'
 import Store from 'electron-store'
-import { getCurrentWindow, app } from '@electron/remote'
+import { getCurrentWindow } from '@electron/remote'
 
 import { install } from './install'
 import OllamaIcon from './ollama.svg'
@@ -19,7 +19,7 @@ export default function () {
   const [step, setStep] = useState<Step>(Step.WELCOME)
   const [commandCopied, setCommandCopied] = useState<boolean>(false)
 
-  const command = 'ollama run llama3.1'
+  const command = 'ollama run llama2'
 
   return (
     <div className='drag'>
@@ -51,15 +51,10 @@ export default function () {
               <div className='mx-auto'>
                 <button
                   onClick={async () => {
-                    try {
-                      await install()
-                      setStep(Step.FINISH)
-                    } catch (e) {
-                      console.error('could not install: ', e)
-                    } finally {
-                      getCurrentWindow().show()
-                      getCurrentWindow().focus()
-                    }
+                    await install()
+                    getCurrentWindow().show()
+                    getCurrentWindow().focus()
+                    setStep(Step.FINISH)
                   }}
                   className='no-drag rounded-dm mx-auto w-[60%] rounded-md bg-black px-4 py-2 text-sm text-white hover:brightness-110'
                 >

app/src/declarations.d.ts

@@ -0,0 +1,4 @@
+declare module '*.svg' {
+  const content: string;
+  export default content;
+}

app/src/index.ts

@@ -1,21 +1,17 @@
-import { spawn, ChildProcess } from 'child_process'
-import { app, autoUpdater, dialog, Tray, Menu, BrowserWindow, MenuItemConstructorOptions, nativeTheme } from 'electron'
+import { spawn } from 'child_process'
+import { app, autoUpdater, dialog, Tray, Menu, BrowserWindow, nativeTheme } from 'electron'
 import Store from 'electron-store'
 import winston from 'winston'
 import 'winston-daily-rotate-file'
 import * as path from 'path'
 
-import { v4 as uuidv4 } from 'uuid'
+import { analytics, id } from './telemetry'
 import { installed } from './install'
 
 require('@electron/remote/main').initialize()
 
-if (require('electron-squirrel-startup')) {
-  app.quit()
-}
-
 const store = new Store()
-
+let tray: Tray | null = null
 let welcomeWindow: BrowserWindow | null = null
 
 declare const MAIN_WINDOW_WEBPACK_ENTRY: string
@@ -32,30 +28,10 @@ const logger = winston.createLogger({
   format: winston.format.printf(info => info.message),
 })
 
-app.on('ready', () => {
-  const gotTheLock = app.requestSingleInstanceLock()
-  if (!gotTheLock) {
-    app.exit(0)
-    return
-  }
-
-  app.on('second-instance', () => {
-    if (app.hasSingleInstanceLock()) {
-      app.releaseSingleInstanceLock()
-    }
-
-    if (proc) {
-      proc.off('exit', restart)
-      proc.kill()
-    }
-
-    app.exit(0)
-  })
-
-  app.focus({ steal: true })
-
-  init()
-})
+const SingleInstanceLock = app.requestSingleInstanceLock()
+if (!SingleInstanceLock) {
+  app.quit()
+}
 
 function firstRunWindow() {
   // Create the browser window.
@@ -71,74 +47,65 @@ function firstRunWindow() {
       nodeIntegration: true,
       contextIsolation: false,
     },
+    alwaysOnTop: true,
   })
 
   require('@electron/remote/main').enable(welcomeWindow.webContents)
 
+  // and load the index.html of the app.
   welcomeWindow.loadURL(MAIN_WINDOW_WEBPACK_ENTRY)
+
   welcomeWindow.on('ready-to-show', () => welcomeWindow.show())
-  welcomeWindow.on('closed', () => {
-    if (process.platform === 'darwin') {
-      app.dock.hide()
+
+  // for debugging
+  // welcomeWindow.webContents.openDevTools()
+
+  if (process.platform === 'darwin') {
+    app.dock.hide()
+  }
+}
+
+function createSystemtray() {
+  let iconPath = nativeTheme.shouldUseDarkColors
+    ? path.join(__dirname, '..', '..', 'assets', 'ollama_icon_16x16Template.png') 
+    : path.join(__dirname, '..', '..', 'assets', 'ollama_outline_icon_16x16Template.png')
+
+  if (app.isPackaged) {
+    iconPath = nativeTheme.shouldUseDarkColors
+    ? path.join(process.resourcesPath, 'ollama_icon_16x16Template.png') 
+    : path.join(process.resourcesPath, 'ollama_outline_icon_16x16Template.png')
+  }
+
+  tray = new Tray(iconPath)
+
+  nativeTheme.on('updated', function theThemeHasChanged () {
+    if (nativeTheme.shouldUseDarkColors) {
+      app.isPackaged 
+        ? tray.setImage(path.join(process.resourcesPath, 'ollama_icon_16x16Template.png')) 
+        : tray.setImage(path.join(__dirname, '..', '..', 'assets', 'ollama_icon_16x16Template.png'))
+    } else {
+      app.isPackaged 
+        ? tray.setImage(path.join(process.resourcesPath, 'ollama_outline_icon_16x16Template.png')) 
+        : tray.setImage(path.join(__dirname, '..', '..', 'assets', 'ollama_outline_icon_16x16Template.png'))
     }
   })
+
+  const contextMenu = Menu.buildFromTemplate([{ role: 'quit', label: 'Quit Ollama', accelerator: 'Command+Q' }])
+
+  tray.setContextMenu(contextMenu)
+  tray.setToolTip('Ollama')
 }
 
-let tray: Tray | null = null
-let updateAvailable = false
-const assetPath = app.isPackaged ? process.resourcesPath : path.join(__dirname, '..', '..', 'assets')
-
-function trayIconPath() {
-  return nativeTheme.shouldUseDarkColors
-    ? updateAvailable
-      ? path.join(assetPath, 'iconDarkUpdateTemplate.png')
-      : path.join(assetPath, 'iconDarkTemplate.png')
-    : updateAvailable
-    ? path.join(assetPath, 'iconUpdateTemplate.png')
-    : path.join(assetPath, 'iconTemplate.png')
+if (require('electron-squirrel-startup')) {
+  app.quit()
 }
 
-function updateTrayIcon() {
-  if (tray) {
-    tray.setImage(trayIconPath())
-  }
-}
-
-function updateTray() {
-  const updateItems: MenuItemConstructorOptions[] = [
-    { label: 'An update is available', enabled: false },
-    {
-      label: 'Restart to update',
-      click: () => autoUpdater.quitAndInstall(),
-    },
-    { type: 'separator' },
-  ]
-
-  const menu = Menu.buildFromTemplate([
-    ...(updateAvailable ? updateItems : []),
-    { role: 'quit', label: 'Quit Ollama', accelerator: 'Command+Q' },
-  ])
-
-  if (!tray) {
-    tray = new Tray(trayIconPath())
-  }
-
-  tray.setToolTip(updateAvailable ? 'An update is available' : 'Ollama')
-  tray.setContextMenu(menu)
-  tray.setImage(trayIconPath())
-
-  nativeTheme.off('updated', updateTrayIcon)
-  nativeTheme.on('updated', updateTrayIcon)
-}
-
-let proc: ChildProcess = null
-
 function server() {
   const binary = app.isPackaged
     ? path.join(process.resourcesPath, 'ollama')
     : path.resolve(process.cwd(), '..', 'ollama')
 
-  proc = spawn(binary, ['serve'])
+  const proc = spawn(binary, ['serve'])
 
   proc.stdout.on('data', data => {
     logger.info(data.toString().trim())
@@ -148,75 +115,24 @@ function server() {
     logger.error(data.toString().trim())
   })
 
+  function restart() {
+    logger.info('Restarting the server...')
+    server()
+  }
+
   proc.on('exit', restart)
-}
 
-function restart() {
-  setTimeout(server, 1000)
-}
-
-app.on('before-quit', () => {
-  if (proc) {
+  app.on('before-quit', () => {
     proc.off('exit', restart)
-    proc.kill('SIGINT') // send SIGINT signal to the server, which also stops any loaded llms
-  }
-})
-
-const updateURL = `https://ollama.com/api/update?os=${process.platform}&arch=${
-  process.arch
-}&version=${app.getVersion()}&id=${id()}`
-
-let latest = ''
-async function isNewReleaseAvailable() {
-  try {
-    const response = await fetch(updateURL)
-
-    if (!response.ok) {
-      return false
-    }
-
-    if (response.status === 204) {
-      return false
-    }
-
-    const data = await response.json()
-
-    const url = data?.url
-    if (!url) {
-      return false
-    }
-
-    if (latest === url) {
-      return false
-    }
-
-    latest = url
-
-    return true
-  } catch (error) {
-    logger.error(`update check failed - ${error}`)
-    return false
-  }
+    proc.kill()
+  })
 }
 
-async function checkUpdate() {
-  const available = await isNewReleaseAvailable()
-  if (available) {
-    logger.info('checking for update')
-    autoUpdater.checkForUpdates()
-  }
+if (process.platform === 'darwin') {
+  app.dock.hide()
 }
 
-function init() {
-  if (app.isPackaged) {
-    checkUpdate()
-    setInterval(() => {
-      checkUpdate()
-    }, 60 * 60 * 1000)
-  }
-
-  updateTray()
-
+app.on('ready', () => {
   if (process.platform === 'darwin') {
     if (app.isPackaged) {
       if (!app.isInApplicationsFolder()) {
@@ -252,13 +168,10 @@ function init() {
     }
   }
 
+  createSystemtray()
   server()
 
   if (store.get('first-time-run') && installed()) {
-    if (process.platform === 'darwin') {
-      app.dock.hide()
-    }
-
     app.setLoginItemSettings({ openAtLogin: app.getLoginItemSettings().openAtLogin })
     return
   }
@@ -266,7 +179,7 @@ function init() {
   // This is the first run or the CLI is no longer installed
   app.setLoginItemSettings({ openAtLogin: true })
   firstRunWindow()
-}
+})
 
 // Quit when all windows are closed, except on macOS. There, it's common
 // for applications and their menu bar to stay active until the user quits
@@ -277,26 +190,45 @@ app.on('window-all-closed', () => {
   }
 })
 
-function id(): string {
-  const id = store.get('id') as string
+// In this file you can include the rest of your app's specific main process
+// code. You can also put them in separate files and import them here.
+autoUpdater.setFeedURL({
+  url: `https://ollama.ai/api/update?os=${process.platform}&arch=${process.arch}&version=${app.getVersion()}`,
+})
 
-  if (id) {
-    return id
-  }
-
-  const uuid = uuidv4()
-  store.set('id', uuid)
-  return uuid
+async function heartbeat() {
+  analytics.track({
+    anonymousId: id(),
+    event: 'heartbeat',
+    properties: {
+      version: app.getVersion(),
+    },
+  })
 }
 
-autoUpdater.setFeedURL({ url: updateURL })
+if (app.isPackaged) {
+  heartbeat()
+  autoUpdater.checkForUpdates()
+  setInterval(() => {
+    heartbeat()
+    autoUpdater.checkForUpdates()
+  }, 60 * 60 * 1000)
+}
 
 autoUpdater.on('error', e => {
   logger.error(`update check failed - ${e.message}`)
-  console.error(`update check failed - ${e.message}`)
 })
 
-autoUpdater.on('update-downloaded', () => {
-  updateAvailable = true
-  updateTray()
+autoUpdater.on('update-downloaded', (event, releaseNotes, releaseName) => {
+  dialog
+    .showMessageBox({
+      type: 'info',
+      buttons: ['Restart Now', 'Later'],
+      title: 'New update available',
+      message: process.platform === 'win32' ? releaseNotes : releaseName,
+      detail: 'A new version of Ollama is available. Restart to apply the update.',
+    })
+    .then(returnValue => {
+      if (returnValue.response === 0) autoUpdater.quitAndInstall()
+    })
 })

app/src/install.ts

@@ -15,7 +15,12 @@ export function installed() {
 export async function install() {
   const command = `do shell script "mkdir -p ${path.dirname(
     symlinkPath
-  )} && ln -F -s \\"${ollama}\\" \\"${symlinkPath}\\"" with administrator privileges`
+  )} && ln -F -s ${ollama} ${symlinkPath}" with administrator privileges`
 
-  await exec(`osascript -e '${command}'`)
+  try {
+    await exec(`osascript -e '${command}'`)
+  } catch (error) {
+    console.error(`cli: failed to install cli: ${error.message}`)
+    return
+  }
 }

app/src/telemetry.ts

@@ -0,0 +1,19 @@
+import { Analytics } from '@segment/analytics-node'
+import { v4 as uuidv4 } from 'uuid'
+import Store from 'electron-store'
+
+const store = new Store()
+
+export const analytics = new Analytics({ writeKey: process.env.TELEMETRY_WRITE_KEY || '<empty>' })
+
+export function id(): string {
+  const id = store.get('id') as string
+
+  if (id) {
+    return id
+  }
+
+  const uuid = uuidv4()
+  store.set('id', uuid)
+  return uuid
+}

app/store/store.go

@@ -1,97 +0,0 @@
-package store
-
-import (
-	"encoding/json"
-	"errors"
-	"fmt"
-	"log/slog"
-	"os"
-	"path/filepath"
-	"sync"
-
-	"github.com/google/uuid"
-)
-
-type Store struct {
-	ID           string `json:"id"`
-	FirstTimeRun bool   `json:"first-time-run"`
-}
-
-var (
-	lock  sync.Mutex
-	store Store
-)
-
-func GetID() string {
-	lock.Lock()
-	defer lock.Unlock()
-	if store.ID == "" {
-		initStore()
-	}
-	return store.ID
-}
-
-func GetFirstTimeRun() bool {
-	lock.Lock()
-	defer lock.Unlock()
-	if store.ID == "" {
-		initStore()
-	}
-	return store.FirstTimeRun
-}
-
-func SetFirstTimeRun(val bool) {
-	lock.Lock()
-	defer lock.Unlock()
-	if store.FirstTimeRun == val {
-		return
-	}
-	store.FirstTimeRun = val
-	writeStore(getStorePath())
-}
-
-// lock must be held
-func initStore() {
-	storeFile, err := os.Open(getStorePath())
-	if err == nil {
-		defer storeFile.Close()
-		err = json.NewDecoder(storeFile).Decode(&store)
-		if err == nil {
-			slog.Debug(fmt.Sprintf("loaded existing store %s - ID: %s", getStorePath(), store.ID))
-			return
-		}
-	} else if !errors.Is(err, os.ErrNotExist) {
-		slog.Debug(fmt.Sprintf("unexpected error searching for store: %s", err))
-	}
-	slog.Debug("initializing new store")
-	store.ID = uuid.New().String()
-	writeStore(getStorePath())
-}
-
-func writeStore(storeFilename string) {
-	ollamaDir := filepath.Dir(storeFilename)
-	_, err := os.Stat(ollamaDir)
-	if errors.Is(err, os.ErrNotExist) {
-		if err := os.MkdirAll(ollamaDir, 0o755); err != nil {
-			slog.Error(fmt.Sprintf("create ollama dir %s: %v", ollamaDir, err))
-			return
-		}
-	}
-	payload, err := json.Marshal(store)
-	if err != nil {
-		slog.Error(fmt.Sprintf("failed to marshal store: %s", err))
-		return
-	}
-	fp, err := os.OpenFile(storeFilename, os.O_WRONLY|os.O_CREATE|os.O_TRUNC, 0o755)
-	if err != nil {
-		slog.Error(fmt.Sprintf("write store payload %s: %v", storeFilename, err))
-		return
-	}
-	defer fp.Close()
-	if n, err := fp.Write(payload); err != nil || n != len(payload) {
-		slog.Error(fmt.Sprintf("write store payload %s: %d vs %d -- %v", storeFilename, n, len(payload), err))
-		return
-	}
-	slog.Debug("Store contents: " + string(payload))
-	slog.Info(fmt.Sprintf("wrote store: %s", storeFilename))
-}

app/store/store_darwin.go

@@ -1,13 +0,0 @@
-package store
-
-import (
-	"os"
-	"path/filepath"
-)
-
-func getStorePath() string {
-	// TODO - system wide location?
-
-	home := os.Getenv("HOME")
-	return filepath.Join(home, "Library", "Application Support", "Ollama", "config.json")
-}

app/store/store_linux.go

@@ -1,16 +0,0 @@
-package store
-
-import (
-	"os"
-	"path/filepath"
-)
-
-func getStorePath() string {
-	if os.Geteuid() == 0 {
-		// TODO where should we store this on linux for system-wide operation?
-		return "/etc/ollama/config.json"
-	}
-
-	home := os.Getenv("HOME")
-	return filepath.Join(home, ".ollama", "config.json")
-}

app/store/store_windows.go

@@ -1,11 +0,0 @@
-package store
-
-import (
-	"os"
-	"path/filepath"
-)
-
-func getStorePath() string {
-	localAppData := os.Getenv("LOCALAPPDATA")
-	return filepath.Join(localAppData, "Ollama", "config.json")
-}

app/tray/commontray/types.go

@@ -1,24 +0,0 @@
-package commontray
-
-var (
-	Title   = "Ollama"
-	ToolTip = "Ollama"
-
-	UpdateIconName = "tray_upgrade"
-	IconName       = "tray"
-)
-
-type Callbacks struct {
-	Quit       chan struct{}
-	Update     chan struct{}
-	DoFirstUse chan struct{}
-	ShowLogs   chan struct{}
-}
-
-type OllamaTray interface {
-	GetCallbacks() Callbacks
-	Run()
-	UpdateAvailable(ver string) error
-	DisplayFirstUseNotification() error
-	Quit()
-}

app/tray/tray.go

@@ -1,28 +0,0 @@
-package tray
-
-import (
-	"fmt"
-	"runtime"
-
-	"github.com/ollama/ollama/app/assets"
-	"github.com/ollama/ollama/app/tray/commontray"
-)
-
-func NewTray() (commontray.OllamaTray, error) {
-	extension := ".png"
-	if runtime.GOOS == "windows" {
-		extension = ".ico"
-	}
-	iconName := commontray.UpdateIconName + extension
-	updateIcon, err := assets.GetIcon(iconName)
-	if err != nil {
-		return nil, fmt.Errorf("failed to load icon %s: %w", iconName, err)
-	}
-	iconName = commontray.IconName + extension
-	icon, err := assets.GetIcon(iconName)
-	if err != nil {
-		return nil, fmt.Errorf("failed to load icon %s: %w", iconName, err)
-	}
-
-	return InitPlatformTray(icon, updateIcon)
-}

app/tray/tray_nonwindows.go

@@ -1,13 +0,0 @@
-//go:build !windows
-
-package tray
-
-import (
-	"fmt"
-
-	"github.com/ollama/ollama/app/tray/commontray"
-)
-
-func InitPlatformTray(icon, updateIcon []byte) (commontray.OllamaTray, error) {
-	return nil, fmt.Errorf("NOT IMPLEMENTED YET")
-}

app/tray/tray_windows.go

@@ -1,10 +0,0 @@
-package tray
-
-import (
-	"github.com/ollama/ollama/app/tray/commontray"
-	"github.com/ollama/ollama/app/tray/wintray"
-)
-
-func InitPlatformTray(icon, updateIcon []byte) (commontray.OllamaTray, error) {
-	return wintray.InitTray(icon, updateIcon)
-}

app/tray/wintray/eventloop.go

@@ -1,183 +0,0 @@
-//go:build windows
-
-package wintray
-
-import (
-	"fmt"
-	"log/slog"
-	"sync"
-	"unsafe"
-
-	"golang.org/x/sys/windows"
-)
-
-var (
-	quitOnce sync.Once
-)
-
-func (t *winTray) Run() {
-	nativeLoop()
-}
-
-func nativeLoop() {
-	// Main message pump.
-	slog.Debug("starting event handling loop")
-	m := &struct {
-		WindowHandle windows.Handle
-		Message      uint32
-		Wparam       uintptr
-		Lparam       uintptr
-		Time         uint32
-		Pt           point
-		LPrivate     uint32
-	}{}
-	for {
-		ret, _, err := pGetMessage.Call(uintptr(unsafe.Pointer(m)), 0, 0, 0)
-
-		// If the function retrieves a message other than WM_QUIT, the return value is nonzero.
-		// If the function retrieves the WM_QUIT message, the return value is zero.
-		// If there is an error, the return value is -1
-		// https://msdn.microsoft.com/en-us/library/windows/desktop/ms644936(v=vs.85).aspx
-		switch int32(ret) {
-		case -1:
-			slog.Error(fmt.Sprintf("get message failure: %v", err))
-			return
-		case 0:
-			return
-		default:
-			pTranslateMessage.Call(uintptr(unsafe.Pointer(m))) //nolint:errcheck
-			pDispatchMessage.Call(uintptr(unsafe.Pointer(m)))  //nolint:errcheck
-		}
-	}
-}
-
-// WindowProc callback function that processes messages sent to a window.
-// https://msdn.microsoft.com/en-us/library/windows/desktop/ms633573(v=vs.85).aspx
-func (t *winTray) wndProc(hWnd windows.Handle, message uint32, wParam, lParam uintptr) (lResult uintptr) {
-	const (
-		WM_RBUTTONUP   = 0x0205
-		WM_LBUTTONUP   = 0x0202
-		WM_COMMAND     = 0x0111
-		WM_ENDSESSION  = 0x0016
-		WM_CLOSE       = 0x0010
-		WM_DESTROY     = 0x0002
-		WM_MOUSEMOVE   = 0x0200
-		WM_LBUTTONDOWN = 0x0201
-	)
-	switch message {
-	case WM_COMMAND:
-		menuItemId := int32(wParam)
-		// https://docs.microsoft.com/en-us/windows/win32/menurc/wm-command#menus
-		switch menuItemId {
-		case quitMenuID:
-			select {
-			case t.callbacks.Quit <- struct{}{}:
-			// should not happen but in case not listening
-			default:
-				slog.Error("no listener on Quit")
-			}
-		case updateMenuID:
-			select {
-			case t.callbacks.Update <- struct{}{}:
-			// should not happen but in case not listening
-			default:
-				slog.Error("no listener on Update")
-			}
-		case diagLogsMenuID:
-			select {
-			case t.callbacks.ShowLogs <- struct{}{}:
-			// should not happen but in case not listening
-			default:
-				slog.Error("no listener on ShowLogs")
-			}
-		default:
-			slog.Debug(fmt.Sprintf("Unexpected menu item id: %d", menuItemId))
-		}
-	case WM_CLOSE:
-		boolRet, _, err := pDestroyWindow.Call(uintptr(t.window))
-		if boolRet == 0 {
-			slog.Error(fmt.Sprintf("failed to destroy window: %s", err))
-		}
-		err = t.wcex.unregister()
-		if err != nil {
-			slog.Error(fmt.Sprintf("failed to uregister windo %s", err))
-		}
-	case WM_DESTROY:
-		// same as WM_ENDSESSION, but throws 0 exit code after all
-		defer pPostQuitMessage.Call(uintptr(int32(0))) //nolint:errcheck
-		fallthrough
-	case WM_ENDSESSION:
-		t.muNID.Lock()
-		if t.nid != nil {
-			err := t.nid.delete()
-			if err != nil {
-				slog.Error(fmt.Sprintf("failed to delete nid: %s", err))
-			}
-		}
-		t.muNID.Unlock()
-	case t.wmSystrayMessage:
-		switch lParam {
-		case WM_MOUSEMOVE, WM_LBUTTONDOWN:
-			// Ignore these...
-		case WM_RBUTTONUP, WM_LBUTTONUP:
-			err := t.showMenu()
-			if err != nil {
-				slog.Error(fmt.Sprintf("failed to show menu: %s", err))
-			}
-		case 0x405: // TODO - how is this magic value derived for the notification left click
-			if t.pendingUpdate {
-				select {
-				case t.callbacks.Update <- struct{}{}:
-				// should not happen but in case not listening
-				default:
-					slog.Error("no listener on Update")
-				}
-			} else {
-				select {
-				case t.callbacks.DoFirstUse <- struct{}{}:
-				// should not happen but in case not listening
-				default:
-					slog.Error("no listener on DoFirstUse")
-				}
-			}
-		case 0x404: // Middle click or close notification
-			// slog.Debug("doing nothing on close of first time notification")
-		default:
-			// 0x402 also seems common - what is it?
-			slog.Debug(fmt.Sprintf("unmanaged app message, lParm: 0x%x", lParam))
-		}
-	case t.wmTaskbarCreated: // on explorer.exe restarts
-		t.muNID.Lock()
-		err := t.nid.add()
-		if err != nil {
-			slog.Error(fmt.Sprintf("failed to refresh the taskbar on explorer restart: %s", err))
-		}
-		t.muNID.Unlock()
-	default:
-		// Calls the default window procedure to provide default processing for any window messages that an application does not process.
-		// https://msdn.microsoft.com/en-us/library/windows/desktop/ms633572(v=vs.85).aspx
-		lResult, _, _ = pDefWindowProc.Call(
-			uintptr(hWnd),
-			uintptr(message),
-			wParam,
-			lParam,
-		)
-	}
-	return
-}
-
-func (t *winTray) Quit() {
-	quitOnce.Do(quit)
-}
-
-func quit() {
-	boolRet, _, err := pPostMessage.Call(
-		uintptr(wt.window),
-		WM_CLOSE,
-		0,
-		0,
-	)
-	if boolRet == 0 {
-		slog.Error(fmt.Sprintf("failed to post close message on shutdown %s", err))
-	}
-}

app/tray/wintray/menus.go

@@ -1,71 +0,0 @@
-//go:build windows
-
-package wintray
-
-import (
-	"fmt"
-	"log/slog"
-	"unsafe"
-
-	"golang.org/x/sys/windows"
-)
-
-const (
-	updatAvailableMenuID = 1
-	updateMenuID         = updatAvailableMenuID + 1
-	separatorMenuID      = updateMenuID + 1
-	diagLogsMenuID       = separatorMenuID + 1
-	diagSeparatorMenuID  = diagLogsMenuID + 1
-	quitMenuID           = diagSeparatorMenuID + 1
-)
-
-func (t *winTray) initMenus() error {
-	if err := t.addOrUpdateMenuItem(diagLogsMenuID, 0, diagLogsMenuTitle, false); err != nil {
-		return fmt.Errorf("unable to create menu entries %w\n", err)
-	}
-	if err := t.addSeparatorMenuItem(diagSeparatorMenuID, 0); err != nil {
-		return fmt.Errorf("unable to create menu entries %w", err)
-	}
-	if err := t.addOrUpdateMenuItem(quitMenuID, 0, quitMenuTitle, false); err != nil {
-		return fmt.Errorf("unable to create menu entries %w\n", err)
-	}
-	return nil
-}
-
-func (t *winTray) UpdateAvailable(ver string) error {
-	if !t.updateNotified {
-		slog.Debug("updating menu and sending notification for new update")
-		if err := t.addOrUpdateMenuItem(updatAvailableMenuID, 0, updateAvailableMenuTitle, true); err != nil {
-			return fmt.Errorf("unable to create menu entries %w", err)
-		}
-		if err := t.addOrUpdateMenuItem(updateMenuID, 0, updateMenutTitle, false); err != nil {
-			return fmt.Errorf("unable to create menu entries %w", err)
-		}
-		if err := t.addSeparatorMenuItem(separatorMenuID, 0); err != nil {
-			return fmt.Errorf("unable to create menu entries %w", err)
-		}
-		iconFilePath, err := iconBytesToFilePath(wt.updateIcon)
-		if err != nil {
-			return fmt.Errorf("unable to write icon data to temp file: %w", err)
-		}
-		if err := wt.setIcon(iconFilePath); err != nil {
-			return fmt.Errorf("unable to set icon: %w", err)
-		}
-		t.updateNotified = true
-
-		t.pendingUpdate = true
-		// Now pop up the notification
-		t.muNID.Lock()
-		defer t.muNID.Unlock()
-		copy(t.nid.InfoTitle[:], windows.StringToUTF16(updateTitle))
-		copy(t.nid.Info[:], windows.StringToUTF16(fmt.Sprintf(updateMessage, ver)))
-		t.nid.Flags |= NIF_INFO
-		t.nid.Timeout = 10
-		t.nid.Size = uint32(unsafe.Sizeof(*wt.nid))
-		err = t.nid.modify()
-		if err != nil {
-			return err
-		}
-	}
-	return nil
-}

app/tray/wintray/messages.go

@@ -1,15 +0,0 @@
-//go:build windows
-
-package wintray
-
-const (
-	firstTimeTitle   = "Ollama is running"
-	firstTimeMessage = "Click here to get started"
-	updateTitle      = "Update available"
-	updateMessage    = "Ollama version %s is ready to install"
-
-	quitMenuTitle            = "Quit Ollama"
-	updateAvailableMenuTitle = "An update is available"
-	updateMenutTitle         = "Restart to update"
-	diagLogsMenuTitle        = "View logs"
-)

app/tray/wintray/notifyicon.go

@@ -1,66 +0,0 @@
-//go:build windows
-
-package wintray
-
-import (
-	"unsafe"
-
-	"golang.org/x/sys/windows"
-)
-
-// Contains information that the system needs to display notifications in the notification area.
-// Used by Shell_NotifyIcon.
-// https://msdn.microsoft.com/en-us/library/windows/desktop/bb773352(v=vs.85).aspx
-// https://msdn.microsoft.com/en-us/library/windows/desktop/bb762159
-type notifyIconData struct {
-	Size                       uint32
-	Wnd                        windows.Handle
-	ID, Flags, CallbackMessage uint32
-	Icon                       windows.Handle
-	Tip                        [128]uint16
-	State, StateMask           uint32
-	Info                       [256]uint16
-	// Timeout, Version           uint32
-	Timeout uint32
-
-	InfoTitle   [64]uint16
-	InfoFlags   uint32
-	GuidItem    windows.GUID
-	BalloonIcon windows.Handle
-}
-
-func (nid *notifyIconData) add() error {
-	const NIM_ADD = 0x00000000
-	res, _, err := pShellNotifyIcon.Call(
-		uintptr(NIM_ADD),
-		uintptr(unsafe.Pointer(nid)),
-	)
-	if res == 0 {
-		return err
-	}
-	return nil
-}
-
-func (nid *notifyIconData) modify() error {
-	const NIM_MODIFY = 0x00000001
-	res, _, err := pShellNotifyIcon.Call(
-		uintptr(NIM_MODIFY),
-		uintptr(unsafe.Pointer(nid)),
-	)
-	if res == 0 {
-		return err
-	}
-	return nil
-}
-
-func (nid *notifyIconData) delete() error {
-	const NIM_DELETE = 0x00000002
-	res, _, err := pShellNotifyIcon.Call(
-		uintptr(NIM_DELETE),
-		uintptr(unsafe.Pointer(nid)),
-	)
-	if res == 0 {
-		return err
-	}
-	return nil
-}

app/tray/wintray/tray.go

@@ -1,481 +0,0 @@
-//go:build windows
-
-package wintray
-
-import (
-	"crypto/md5"
-	"encoding/hex"
-	"fmt"
-	"log/slog"
-	"os"
-	"path/filepath"
-	"sort"
-	"sync"
-	"unsafe"
-
-	"github.com/ollama/ollama/app/tray/commontray"
-	"golang.org/x/sys/windows"
-)
-
-// Helpful sources: https://github.com/golang/exp/blob/master/shiny/driver/internal/win32
-
-// Contains information about loaded resources
-type winTray struct {
-	instance,
-	icon,
-	cursor,
-	window windows.Handle
-
-	loadedImages   map[string]windows.Handle
-	muLoadedImages sync.RWMutex
-
-	// menus keeps track of the submenus keyed by the menu item ID, plus 0
-	// which corresponds to the main popup menu.
-	menus    map[uint32]windows.Handle
-	muMenus  sync.RWMutex
-	menuOf   map[uint32]windows.Handle
-	muMenuOf sync.RWMutex
-	// menuItemIcons maintains the bitmap of each menu item (if applies). It's
-	// needed to show the icon correctly when showing a previously hidden menu
-	// item again.
-	// menuItemIcons   map[uint32]windows.Handle
-	// muMenuItemIcons sync.RWMutex
-	visibleItems   map[uint32][]uint32
-	muVisibleItems sync.RWMutex
-
-	nid   *notifyIconData
-	muNID sync.RWMutex
-	wcex  *wndClassEx
-
-	wmSystrayMessage,
-	wmTaskbarCreated uint32
-
-	pendingUpdate  bool
-	updateNotified bool // Only pop up the notification once - TODO consider daily nag?
-	// Callbacks
-	callbacks  commontray.Callbacks
-	normalIcon []byte
-	updateIcon []byte
-}
-
-var wt winTray
-
-func (t *winTray) GetCallbacks() commontray.Callbacks {
-	return t.callbacks
-}
-
-func InitTray(icon, updateIcon []byte) (*winTray, error) {
-	wt.callbacks.Quit = make(chan struct{})
-	wt.callbacks.Update = make(chan struct{})
-	wt.callbacks.ShowLogs = make(chan struct{})
-	wt.callbacks.DoFirstUse = make(chan struct{})
-	wt.normalIcon = icon
-	wt.updateIcon = updateIcon
-	if err := wt.initInstance(); err != nil {
-		return nil, fmt.Errorf("Unable to init instance: %w\n", err)
-	}
-
-	if err := wt.createMenu(); err != nil {
-		return nil, fmt.Errorf("Unable to create menu: %w\n", err)
-	}
-
-	iconFilePath, err := iconBytesToFilePath(wt.normalIcon)
-	if err != nil {
-		return nil, fmt.Errorf("Unable to write icon data to temp file: %w", err)
-	}
-	if err := wt.setIcon(iconFilePath); err != nil {
-		return nil, fmt.Errorf("Unable to set icon: %w", err)
-	}
-
-	return &wt, wt.initMenus()
-}
-
-func (t *winTray) initInstance() error {
-	const (
-		className  = "OllamaClass"
-		windowName = ""
-	)
-
-	t.wmSystrayMessage = WM_USER + 1
-	t.visibleItems = make(map[uint32][]uint32)
-	t.menus = make(map[uint32]windows.Handle)
-	t.menuOf = make(map[uint32]windows.Handle)
-
-	t.loadedImages = make(map[string]windows.Handle)
-
-	taskbarEventNamePtr, _ := windows.UTF16PtrFromString("TaskbarCreated")
-	// https://msdn.microsoft.com/en-us/library/windows/desktop/ms644947
-	res, _, err := pRegisterWindowMessage.Call(
-		uintptr(unsafe.Pointer(taskbarEventNamePtr)),
-	)
-	if res == 0 { // success 0xc000-0xfff
-		return fmt.Errorf("failed to register window: %w", err)
-	}
-	t.wmTaskbarCreated = uint32(res)
-
-	instanceHandle, _, err := pGetModuleHandle.Call(0)
-	if instanceHandle == 0 {
-		return err
-	}
-	t.instance = windows.Handle(instanceHandle)
-
-	// https://msdn.microsoft.com/en-us/library/windows/desktop/ms648072(v=vs.85).aspx
-	iconHandle, _, err := pLoadIcon.Call(0, uintptr(IDI_APPLICATION))
-	if iconHandle == 0 {
-		return err
-	}
-	t.icon = windows.Handle(iconHandle)
-
-	// https://msdn.microsoft.com/en-us/library/windows/desktop/ms648391(v=vs.85).aspx
-	cursorHandle, _, err := pLoadCursor.Call(0, uintptr(IDC_ARROW))
-	if cursorHandle == 0 {
-		return err
-	}
-	t.cursor = windows.Handle(cursorHandle)
-
-	classNamePtr, err := windows.UTF16PtrFromString(className)
-	if err != nil {
-		return err
-	}
-
-	windowNamePtr, err := windows.UTF16PtrFromString(windowName)
-	if err != nil {
-		return err
-	}
-
-	t.wcex = &wndClassEx{
-		Style:      CS_HREDRAW | CS_VREDRAW,
-		WndProc:    windows.NewCallback(t.wndProc),
-		Instance:   t.instance,
-		Icon:       t.icon,
-		Cursor:     t.cursor,
-		Background: windows.Handle(6), // (COLOR_WINDOW + 1)
-		ClassName:  classNamePtr,
-		IconSm:     t.icon,
-	}
-	if err := t.wcex.register(); err != nil {
-		return err
-	}
-
-	windowHandle, _, err := pCreateWindowEx.Call(
-		uintptr(0),
-		uintptr(unsafe.Pointer(classNamePtr)),
-		uintptr(unsafe.Pointer(windowNamePtr)),
-		uintptr(WS_OVERLAPPEDWINDOW),
-		uintptr(CW_USEDEFAULT),
-		uintptr(CW_USEDEFAULT),
-		uintptr(CW_USEDEFAULT),
-		uintptr(CW_USEDEFAULT),
-		uintptr(0),
-		uintptr(0),
-		uintptr(t.instance),
-		uintptr(0),
-	)
-	if windowHandle == 0 {
-		return err
-	}
-	t.window = windows.Handle(windowHandle)
-
-	pShowWindow.Call(uintptr(t.window), uintptr(SW_HIDE)) //nolint:errcheck
-
-	boolRet, _, err := pUpdateWindow.Call(uintptr(t.window))
-	if boolRet == 0 {
-		slog.Error(fmt.Sprintf("failed to update window: %s", err))
-	}
-
-	t.muNID.Lock()
-	defer t.muNID.Unlock()
-	t.nid = &notifyIconData{
-		Wnd:             t.window,
-		ID:              100,
-		Flags:           NIF_MESSAGE,
-		CallbackMessage: t.wmSystrayMessage,
-	}
-	t.nid.Size = uint32(unsafe.Sizeof(*t.nid))
-
-	return t.nid.add()
-}
-
-func (t *winTray) createMenu() error {
-	menuHandle, _, err := pCreatePopupMenu.Call()
-	if menuHandle == 0 {
-		return err
-	}
-	t.menus[0] = windows.Handle(menuHandle)
-
-	// https://msdn.microsoft.com/en-us/library/windows/desktop/ms647575(v=vs.85).aspx
-	mi := struct {
-		Size, Mask, Style, Max uint32
-		Background             windows.Handle
-		ContextHelpID          uint32
-		MenuData               uintptr
-	}{
-		Mask: MIM_APPLYTOSUBMENUS,
-	}
-	mi.Size = uint32(unsafe.Sizeof(mi))
-
-	res, _, err := pSetMenuInfo.Call(
-		uintptr(t.menus[0]),
-		uintptr(unsafe.Pointer(&mi)),
-	)
-	if res == 0 {
-		return err
-	}
-	return nil
-}
-
-// Contains information about a menu item.
-// https://msdn.microsoft.com/en-us/library/windows/desktop/ms647578(v=vs.85).aspx
-type menuItemInfo struct {
-	Size, Mask, Type, State     uint32
-	ID                          uint32
-	SubMenu, Checked, Unchecked windows.Handle
-	ItemData                    uintptr
-	TypeData                    *uint16
-	Cch                         uint32
-	BMPItem                     windows.Handle
-}
-
-func (t *winTray) addOrUpdateMenuItem(menuItemId uint32, parentId uint32, title string, disabled bool) error {
-	titlePtr, err := windows.UTF16PtrFromString(title)
-	if err != nil {
-		return err
-	}
-
-	mi := menuItemInfo{
-		Mask:     MIIM_FTYPE | MIIM_STRING | MIIM_ID | MIIM_STATE,
-		Type:     MFT_STRING,
-		ID:       menuItemId,
-		TypeData: titlePtr,
-		Cch:      uint32(len(title)),
-	}
-	mi.Size = uint32(unsafe.Sizeof(mi))
-	if disabled {
-		mi.State |= MFS_DISABLED
-	}
-
-	var res uintptr
-	t.muMenus.RLock()
-	menu := t.menus[parentId]
-	t.muMenus.RUnlock()
-	if t.getVisibleItemIndex(parentId, menuItemId) != -1 {
-		// We set the menu item info based on the menuID
-		boolRet, _, err := pSetMenuItemInfo.Call(
-			uintptr(menu),
-			uintptr(menuItemId),
-			0,
-			uintptr(unsafe.Pointer(&mi)),
-		)
-		if boolRet == 0 {
-			return fmt.Errorf("failed to set menu item: %w", err)
-		}
-	}
-
-	if res == 0 {
-		// Menu item does not already exist, create it
-		t.muMenus.RLock()
-		submenu, exists := t.menus[menuItemId]
-		t.muMenus.RUnlock()
-		if exists {
-			mi.Mask |= MIIM_SUBMENU
-			mi.SubMenu = submenu
-		}
-		t.addToVisibleItems(parentId, menuItemId)
-		position := t.getVisibleItemIndex(parentId, menuItemId)
-		res, _, err = pInsertMenuItem.Call(
-			uintptr(menu),
-			uintptr(position),
-			1,
-			uintptr(unsafe.Pointer(&mi)),
-		)
-		if res == 0 {
-			t.delFromVisibleItems(parentId, menuItemId)
-			return err
-		}
-		t.muMenuOf.Lock()
-		t.menuOf[menuItemId] = menu
-		t.muMenuOf.Unlock()
-	}
-
-	return nil
-}
-
-func (t *winTray) addSeparatorMenuItem(menuItemId, parentId uint32) error {
-	mi := menuItemInfo{
-		Mask: MIIM_FTYPE | MIIM_ID | MIIM_STATE,
-		Type: MFT_SEPARATOR,
-		ID:   menuItemId,
-	}
-
-	mi.Size = uint32(unsafe.Sizeof(mi))
-
-	t.addToVisibleItems(parentId, menuItemId)
-	position := t.getVisibleItemIndex(parentId, menuItemId)
-	t.muMenus.RLock()
-	menu := uintptr(t.menus[parentId])
-	t.muMenus.RUnlock()
-	res, _, err := pInsertMenuItem.Call(
-		menu,
-		uintptr(position),
-		1,
-		uintptr(unsafe.Pointer(&mi)),
-	)
-	if res == 0 {
-		return err
-	}
-
-	return nil
-}
-
-// func (t *winTray) hideMenuItem(menuItemId, parentId uint32) error {
-// 	const ERROR_SUCCESS syscall.Errno = 0
-
-// 	t.muMenus.RLock()
-// 	menu := uintptr(t.menus[parentId])
-// 	t.muMenus.RUnlock()
-// 	res, _, err := pRemoveMenu.Call(
-// 		menu,
-// 		uintptr(menuItemId),
-// 		MF_BYCOMMAND,
-// 	)
-// 	if res == 0 && err.(syscall.Errno) != ERROR_SUCCESS {
-// 		return err
-// 	}
-// 	t.delFromVisibleItems(parentId, menuItemId)
-
-// 	return nil
-// }
-
-func (t *winTray) showMenu() error {
-	p := point{}
-	boolRet, _, err := pGetCursorPos.Call(uintptr(unsafe.Pointer(&p)))
-	if boolRet == 0 {
-		return err
-	}
-	boolRet, _, err = pSetForegroundWindow.Call(uintptr(t.window))
-	if boolRet == 0 {
-		slog.Warn(fmt.Sprintf("failed to bring menu to foreground: %s", err))
-	}
-
-	boolRet, _, err = pTrackPopupMenu.Call(
-		uintptr(t.menus[0]),
-		TPM_BOTTOMALIGN|TPM_LEFTALIGN,
-		uintptr(p.X),
-		uintptr(p.Y),
-		0,
-		uintptr(t.window),
-		0,
-	)
-	if boolRet == 0 {
-		return err
-	}
-
-	return nil
-}
-
-func (t *winTray) delFromVisibleItems(parent, val uint32) {
-	t.muVisibleItems.Lock()
-	defer t.muVisibleItems.Unlock()
-	visibleItems := t.visibleItems[parent]
-	for i, itemval := range visibleItems {
-		if val == itemval {
-			t.visibleItems[parent] = append(visibleItems[:i], visibleItems[i+1:]...)
-			break
-		}
-	}
-}
-
-func (t *winTray) addToVisibleItems(parent, val uint32) {
-	t.muVisibleItems.Lock()
-	defer t.muVisibleItems.Unlock()
-	if visibleItems, exists := t.visibleItems[parent]; !exists {
-		t.visibleItems[parent] = []uint32{val}
-	} else {
-		newvisible := append(visibleItems, val)
-		sort.Slice(newvisible, func(i, j int) bool { return newvisible[i] < newvisible[j] })
-		t.visibleItems[parent] = newvisible
-	}
-}
-
-func (t *winTray) getVisibleItemIndex(parent, val uint32) int {
-	t.muVisibleItems.RLock()
-	defer t.muVisibleItems.RUnlock()
-	for i, itemval := range t.visibleItems[parent] {
-		if val == itemval {
-			return i
-		}
-	}
-	return -1
-}
-
-func iconBytesToFilePath(iconBytes []byte) (string, error) {
-	bh := md5.Sum(iconBytes)
-	dataHash := hex.EncodeToString(bh[:])
-	iconFilePath := filepath.Join(os.TempDir(), "ollama_temp_icon_"+dataHash)
-
-	if _, err := os.Stat(iconFilePath); os.IsNotExist(err) {
-		if err := os.WriteFile(iconFilePath, iconBytes, 0644); err != nil {
-			return "", err
-		}
-	}
-	return iconFilePath, nil
-}
-
-// Loads an image from file and shows it in tray.
-// Shell_NotifyIcon: https://msdn.microsoft.com/en-us/library/windows/desktop/bb762159(v=vs.85).aspx
-func (t *winTray) setIcon(src string) error {
-	h, err := t.loadIconFrom(src)
-	if err != nil {
-		return err
-	}
-
-	t.muNID.Lock()
-	defer t.muNID.Unlock()
-	t.nid.Icon = h
-	t.nid.Flags |= NIF_ICON
-	t.nid.Size = uint32(unsafe.Sizeof(*t.nid))
-
-	return t.nid.modify()
-}
-
-// Loads an image from file to be shown in tray or menu item.
-// LoadImage: https://msdn.microsoft.com/en-us/library/windows/desktop/ms648045(v=vs.85).aspx
-func (t *winTray) loadIconFrom(src string) (windows.Handle, error) {
-	// Save and reuse handles of loaded images
-	t.muLoadedImages.RLock()
-	h, ok := t.loadedImages[src]
-	t.muLoadedImages.RUnlock()
-	if !ok {
-		srcPtr, err := windows.UTF16PtrFromString(src)
-		if err != nil {
-			return 0, err
-		}
-		res, _, err := pLoadImage.Call(
-			0,
-			uintptr(unsafe.Pointer(srcPtr)),
-			IMAGE_ICON,
-			0,
-			0,
-			LR_LOADFROMFILE|LR_DEFAULTSIZE,
-		)
-		if res == 0 {
-			return 0, err
-		}
-		h = windows.Handle(res)
-		t.muLoadedImages.Lock()
-		t.loadedImages[src] = h
-		t.muLoadedImages.Unlock()
-	}
-	return h, nil
-}
-
-func (t *winTray) DisplayFirstUseNotification() error {
-	t.muNID.Lock()
-	defer t.muNID.Unlock()
-	copy(t.nid.InfoTitle[:], windows.StringToUTF16(firstTimeTitle))
-	copy(t.nid.Info[:], windows.StringToUTF16(firstTimeMessage))
-	t.nid.Flags |= NIF_INFO
-	t.nid.Size = uint32(unsafe.Sizeof(*wt.nid))
-
-	return t.nid.modify()
-}

app/tray/wintray/w32api.go

@@ -1,89 +0,0 @@
-//go:build windows
-
-package wintray
-
-import (
-	"runtime"
-
-	"golang.org/x/sys/windows"
-)
-
-var (
-	k32 = windows.NewLazySystemDLL("Kernel32.dll")
-	u32 = windows.NewLazySystemDLL("User32.dll")
-	s32 = windows.NewLazySystemDLL("Shell32.dll")
-
-	pCreatePopupMenu       = u32.NewProc("CreatePopupMenu")
-	pCreateWindowEx        = u32.NewProc("CreateWindowExW")
-	pDefWindowProc         = u32.NewProc("DefWindowProcW")
-	pDestroyWindow         = u32.NewProc("DestroyWindow")
-	pDispatchMessage       = u32.NewProc("DispatchMessageW")
-	pGetCursorPos          = u32.NewProc("GetCursorPos")
-	pGetMessage            = u32.NewProc("GetMessageW")
-	pGetModuleHandle       = k32.NewProc("GetModuleHandleW")
-	pInsertMenuItem        = u32.NewProc("InsertMenuItemW")
-	pLoadCursor            = u32.NewProc("LoadCursorW")
-	pLoadIcon              = u32.NewProc("LoadIconW")
-	pLoadImage             = u32.NewProc("LoadImageW")
-	pPostMessage           = u32.NewProc("PostMessageW")
-	pPostQuitMessage       = u32.NewProc("PostQuitMessage")
-	pRegisterClass         = u32.NewProc("RegisterClassExW")
-	pRegisterWindowMessage = u32.NewProc("RegisterWindowMessageW")
-	pSetForegroundWindow   = u32.NewProc("SetForegroundWindow")
-	pSetMenuInfo           = u32.NewProc("SetMenuInfo")
-	pSetMenuItemInfo       = u32.NewProc("SetMenuItemInfoW")
-	pShellNotifyIcon       = s32.NewProc("Shell_NotifyIconW")
-	pShowWindow            = u32.NewProc("ShowWindow")
-	pTrackPopupMenu        = u32.NewProc("TrackPopupMenu")
-	pTranslateMessage      = u32.NewProc("TranslateMessage")
-	pUnregisterClass       = u32.NewProc("UnregisterClassW")
-	pUpdateWindow          = u32.NewProc("UpdateWindow")
-)
-
-const (
-	CS_HREDRAW          = 0x0002
-	CS_VREDRAW          = 0x0001
-	CW_USEDEFAULT       = 0x80000000
-	IDC_ARROW           = 32512 // Standard arrow
-	IDI_APPLICATION     = 32512
-	IMAGE_ICON          = 1          // Loads an icon
-	LR_DEFAULTSIZE      = 0x00000040 // Loads default-size icon for windows(SM_CXICON x SM_CYICON) if cx, cy are set to zero
-	LR_LOADFROMFILE     = 0x00000010 // Loads the stand-alone image from the file
-	MF_BYCOMMAND        = 0x00000000
-	MFS_DISABLED        = 0x00000003
-	MFT_SEPARATOR       = 0x00000800
-	MFT_STRING          = 0x00000000
-	MIIM_BITMAP         = 0x00000080
-	MIIM_FTYPE          = 0x00000100
-	MIIM_ID             = 0x00000002
-	MIIM_STATE          = 0x00000001
-	MIIM_STRING         = 0x00000040
-	MIIM_SUBMENU        = 0x00000004
-	MIM_APPLYTOSUBMENUS = 0x80000000
-	NIF_ICON            = 0x00000002
-	NIF_INFO            = 0x00000010
-	NIF_MESSAGE         = 0x00000001
-	SW_HIDE             = 0
-	TPM_BOTTOMALIGN     = 0x0020
-	TPM_LEFTALIGN       = 0x0000
-	WM_CLOSE            = 0x0010
-	WM_USER             = 0x0400
-	WS_CAPTION          = 0x00C00000
-	WS_MAXIMIZEBOX      = 0x00010000
-	WS_MINIMIZEBOX      = 0x00020000
-	WS_OVERLAPPED       = 0x00000000
-	WS_OVERLAPPEDWINDOW = WS_OVERLAPPED | WS_CAPTION | WS_SYSMENU | WS_THICKFRAME | WS_MINIMIZEBOX | WS_MAXIMIZEBOX
-	WS_SYSMENU          = 0x00080000
-	WS_THICKFRAME       = 0x00040000
-)
-
-// Not sure if this is actually needed on windows
-func init() {
-	runtime.LockOSThread()
-}
-
-// The POINT structure defines the x- and y- coordinates of a point.
-// https://msdn.microsoft.com/en-us/library/windows/desktop/dd162805(v=vs.85).aspx
-type point struct {
-	X, Y int32
-}

app/tray/wintray/winclass.go

@@ -1,45 +0,0 @@
-//go:build windows
-
-package wintray
-
-import (
-	"unsafe"
-
-	"golang.org/x/sys/windows"
-)
-
-// Contains window class information.
-// It is used with the RegisterClassEx and GetClassInfoEx functions.
-// https://msdn.microsoft.com/en-us/library/ms633577.aspx
-type wndClassEx struct {
-	Size, Style                        uint32
-	WndProc                            uintptr
-	ClsExtra, WndExtra                 int32
-	Instance, Icon, Cursor, Background windows.Handle
-	MenuName, ClassName                *uint16
-	IconSm                             windows.Handle
-}
-
-// Registers a window class for subsequent use in calls to the CreateWindow or CreateWindowEx function.
-// https://msdn.microsoft.com/en-us/library/ms633587.aspx
-func (w *wndClassEx) register() error {
-	w.Size = uint32(unsafe.Sizeof(*w))
-	res, _, err := pRegisterClass.Call(uintptr(unsafe.Pointer(w)))
-	if res == 0 {
-		return err
-	}
-	return nil
-}
-
-// Unregisters a window class, freeing the memory required for the class.
-// https://msdn.microsoft.com/en-us/library/ms644899.aspx
-func (w *wndClassEx) unregister() error {
-	res, _, err := pUnregisterClass.Call(
-		uintptr(unsafe.Pointer(w.ClassName)),
-		uintptr(w.Instance),
-	)
-	if res == 0 {
-		return err
-	}
-	return nil
-}

auth/auth.go

@@ -1,91 +0,0 @@
-package auth
-
-import (
-	"bytes"
-	"context"
-	"crypto/rand"
-	"encoding/base64"
-	"fmt"
-	"io"
-	"log/slog"
-	"os"
-	"path/filepath"
-	"strings"
-
-	"golang.org/x/crypto/ssh"
-)
-
-const defaultPrivateKey = "id_ed25519"
-
-func keyPath() (string, error) {
-	home, err := os.UserHomeDir()
-	if err != nil {
-		return "", err
-	}
-
-	return filepath.Join(home, ".ollama", defaultPrivateKey), nil
-}
-
-func GetPublicKey() (string, error) {
-	keyPath, err := keyPath()
-	if err != nil {
-		return "", err
-	}
-
-	privateKeyFile, err := os.ReadFile(keyPath)
-	if err != nil {
-		slog.Info(fmt.Sprintf("Failed to load private key: %v", err))
-		return "", err
-	}
-
-	privateKey, err := ssh.ParsePrivateKey(privateKeyFile)
-	if err != nil {
-		return "", err
-	}
-
-	publicKey := ssh.MarshalAuthorizedKey(privateKey.PublicKey())
-
-	return strings.TrimSpace(string(publicKey)), nil
-}
-
-func NewNonce(r io.Reader, length int) (string, error) {
-	nonce := make([]byte, length)
-	if _, err := io.ReadFull(r, nonce); err != nil {
-		return "", err
-	}
-
-	return base64.RawURLEncoding.EncodeToString(nonce), nil
-}
-
-func Sign(ctx context.Context, bts []byte) (string, error) {
-	keyPath, err := keyPath()
-	if err != nil {
-		return "", err
-	}
-
-	privateKeyFile, err := os.ReadFile(keyPath)
-	if err != nil {
-		slog.Info(fmt.Sprintf("Failed to load private key: %v", err))
-		return "", err
-	}
-
-	privateKey, err := ssh.ParsePrivateKey(privateKeyFile)
-	if err != nil {
-		return "", err
-	}
-
-	// get the pubkey, but remove the type
-	publicKey := ssh.MarshalAuthorizedKey(privateKey.PublicKey())
-	parts := bytes.Split(publicKey, []byte(" "))
-	if len(parts) < 2 {
-		return "", fmt.Errorf("malformed public key")
-	}
-
-	signedData, err := privateKey.Sign(rand.Reader, bts)
-	if err != nil {
-		return "", err
-	}
-
-	// signature is <pubkey>:<signature>
-	return fmt.Sprintf("%s:%s", bytes.TrimSpace(parts[1]), base64.StdEncoding.EncodeToString(signedData.Blob)), nil
-}

cmd/interactive.go

@@ -1,622 +0,0 @@
-package cmd
-
-import (
-	"cmp"
-	"errors"
-	"fmt"
-	"io"
-	"net/http"
-	"os"
-	"path/filepath"
-	"regexp"
-	"slices"
-	"strings"
-
-	"github.com/spf13/cobra"
-	"golang.org/x/exp/maps"
-
-	"github.com/ollama/ollama/api"
-	"github.com/ollama/ollama/envconfig"
-	"github.com/ollama/ollama/parser"
-	"github.com/ollama/ollama/progress"
-	"github.com/ollama/ollama/readline"
-	"github.com/ollama/ollama/types/errtypes"
-)
-
-type MultilineState int
-
-const (
-	MultilineNone MultilineState = iota
-	MultilinePrompt
-	MultilineSystem
-)
-
-func loadModel(cmd *cobra.Command, opts *runOptions) error {
-	p := progress.NewProgress(os.Stderr)
-	defer p.StopAndClear()
-
-	spinner := progress.NewSpinner("")
-	p.Add("", spinner)
-
-	client, err := api.ClientFromEnvironment()
-	if err != nil {
-		return err
-	}
-
-	chatReq := &api.ChatRequest{
-		Model:     opts.Model,
-		KeepAlive: opts.KeepAlive,
-	}
-
-	return client.Chat(cmd.Context(), chatReq, func(api.ChatResponse) error { return nil })
-}
-
-func generateInteractive(cmd *cobra.Command, opts runOptions) error {
-	usage := func() {
-		fmt.Fprintln(os.Stderr, "Available Commands:")
-		fmt.Fprintln(os.Stderr, "  /set            Set session variables")
-		fmt.Fprintln(os.Stderr, "  /show           Show model information")
-		fmt.Fprintln(os.Stderr, "  /load <model>   Load a session or model")
-		fmt.Fprintln(os.Stderr, "  /save <model>   Save your current session")
-		fmt.Fprintln(os.Stderr, "  /clear          Clear session context")
-		fmt.Fprintln(os.Stderr, "  /bye            Exit")
-		fmt.Fprintln(os.Stderr, "  /?, /help       Help for a command")
-		fmt.Fprintln(os.Stderr, "  /? shortcuts    Help for keyboard shortcuts")
-		fmt.Fprintln(os.Stderr, "")
-		fmt.Fprintln(os.Stderr, "Use \"\"\" to begin a multi-line message.")
-
-		if opts.MultiModal {
-			fmt.Fprintf(os.Stderr, "Use %s to include .jpg or .png images.\n", filepath.FromSlash("/path/to/file"))
-		}
-
-		fmt.Fprintln(os.Stderr, "")
-	}
-
-	usageSet := func() {
-		fmt.Fprintln(os.Stderr, "Available Commands:")
-		fmt.Fprintln(os.Stderr, "  /set parameter ...     Set a parameter")
-		fmt.Fprintln(os.Stderr, "  /set system <string>   Set system message")
-		fmt.Fprintln(os.Stderr, "  /set history           Enable history")
-		fmt.Fprintln(os.Stderr, "  /set nohistory         Disable history")
-		fmt.Fprintln(os.Stderr, "  /set wordwrap          Enable wordwrap")
-		fmt.Fprintln(os.Stderr, "  /set nowordwrap        Disable wordwrap")
-		fmt.Fprintln(os.Stderr, "  /set format json       Enable JSON mode")
-		fmt.Fprintln(os.Stderr, "  /set noformat          Disable formatting")
-		fmt.Fprintln(os.Stderr, "  /set verbose           Show LLM stats")
-		fmt.Fprintln(os.Stderr, "  /set quiet             Disable LLM stats")
-		fmt.Fprintln(os.Stderr, "")
-	}
-
-	usageShortcuts := func() {
-		fmt.Fprintln(os.Stderr, "Available keyboard shortcuts:")
-		fmt.Fprintln(os.Stderr, "  Ctrl + a            Move to the beginning of the line (Home)")
-		fmt.Fprintln(os.Stderr, "  Ctrl + e            Move to the end of the line (End)")
-		fmt.Fprintln(os.Stderr, "   Alt + b            Move back (left) one word")
-		fmt.Fprintln(os.Stderr, "   Alt + f            Move forward (right) one word")
-		fmt.Fprintln(os.Stderr, "  Ctrl + k            Delete the sentence after the cursor")
-		fmt.Fprintln(os.Stderr, "  Ctrl + u            Delete the sentence before the cursor")
-		fmt.Fprintln(os.Stderr, "  Ctrl + w            Delete the word before the cursor")
-		fmt.Fprintln(os.Stderr, "")
-		fmt.Fprintln(os.Stderr, "  Ctrl + l            Clear the screen")
-		fmt.Fprintln(os.Stderr, "  Ctrl + c            Stop the model from responding")
-		fmt.Fprintln(os.Stderr, "  Ctrl + d            Exit ollama (/bye)")
-		fmt.Fprintln(os.Stderr, "")
-	}
-
-	usageShow := func() {
-		fmt.Fprintln(os.Stderr, "Available Commands:")
-		fmt.Fprintln(os.Stderr, "  /show info         Show details for this model")
-		fmt.Fprintln(os.Stderr, "  /show license      Show model license")
-		fmt.Fprintln(os.Stderr, "  /show modelfile    Show Modelfile for this model")
-		fmt.Fprintln(os.Stderr, "  /show parameters   Show parameters for this model")
-		fmt.Fprintln(os.Stderr, "  /show system       Show system message")
-		fmt.Fprintln(os.Stderr, "  /show template     Show prompt template")
-		fmt.Fprintln(os.Stderr, "")
-	}
-
-	// only list out the most common parameters
-	usageParameters := func() {
-		fmt.Fprintln(os.Stderr, "Available Parameters:")
-		fmt.Fprintln(os.Stderr, "  /set parameter seed <int>             Random number seed")
-		fmt.Fprintln(os.Stderr, "  /set parameter num_predict <int>      Max number of tokens to predict")
-		fmt.Fprintln(os.Stderr, "  /set parameter top_k <int>            Pick from top k num of tokens")
-		fmt.Fprintln(os.Stderr, "  /set parameter top_p <float>          Pick token based on sum of probabilities")
-		fmt.Fprintln(os.Stderr, "  /set parameter min_p <float>          Pick token based on top token probability * min_p")
-		fmt.Fprintln(os.Stderr, "  /set parameter num_ctx <int>          Set the context size")
-		fmt.Fprintln(os.Stderr, "  /set parameter temperature <float>    Set creativity level")
-		fmt.Fprintln(os.Stderr, "  /set parameter repeat_penalty <float> How strongly to penalize repetitions")
-		fmt.Fprintln(os.Stderr, "  /set parameter repeat_last_n <int>    Set how far back to look for repetitions")
-		fmt.Fprintln(os.Stderr, "  /set parameter num_gpu <int>          The number of layers to send to the GPU")
-		fmt.Fprintln(os.Stderr, "  /set parameter stop <string> <string> ...   Set the stop parameters")
-		fmt.Fprintln(os.Stderr, "")
-	}
-
-	scanner, err := readline.New(readline.Prompt{
-		Prompt:         ">>> ",
-		AltPrompt:      "... ",
-		Placeholder:    "Send a message (/? for help)",
-		AltPlaceholder: `Use """ to end multi-line input`,
-	})
-	if err != nil {
-		return err
-	}
-
-	if envconfig.NoHistory() {
-		scanner.HistoryDisable()
-	}
-
-	fmt.Print(readline.StartBracketedPaste)
-	defer fmt.Printf(readline.EndBracketedPaste)
-
-	var sb strings.Builder
-	var multiline MultilineState
-
-	for {
-		line, err := scanner.Readline()
-		switch {
-		case errors.Is(err, io.EOF):
-			fmt.Println()
-			return nil
-		case errors.Is(err, readline.ErrInterrupt):
-			if line == "" {
-				fmt.Println("\nUse Ctrl + d or /bye to exit.")
-			}
-
-			scanner.Prompt.UseAlt = false
-			sb.Reset()
-
-			continue
-		case err != nil:
-			return err
-		}
-
-		switch {
-		case multiline != MultilineNone:
-			// check if there's a multiline terminating string
-			before, ok := strings.CutSuffix(line, `"""`)
-			sb.WriteString(before)
-			if !ok {
-				fmt.Fprintln(&sb)
-				continue
-			}
-
-			switch multiline {
-			case MultilineSystem:
-				opts.System = sb.String()
-				opts.Messages = append(opts.Messages, api.Message{Role: "system", Content: opts.System})
-				fmt.Println("Set system message.")
-				sb.Reset()
-			}
-
-			multiline = MultilineNone
-			scanner.Prompt.UseAlt = false
-		case strings.HasPrefix(line, `"""`):
-			line := strings.TrimPrefix(line, `"""`)
-			line, ok := strings.CutSuffix(line, `"""`)
-			sb.WriteString(line)
-			if !ok {
-				// no multiline terminating string; need more input
-				fmt.Fprintln(&sb)
-				multiline = MultilinePrompt
-				scanner.Prompt.UseAlt = true
-			}
-		case scanner.Pasting:
-			fmt.Fprintln(&sb, line)
-			continue
-		case strings.HasPrefix(line, "/list"):
-			args := strings.Fields(line)
-			if err := ListHandler(cmd, args[1:]); err != nil {
-				return err
-			}
-		case strings.HasPrefix(line, "/load"):
-			args := strings.Fields(line)
-			if len(args) != 2 {
-				fmt.Println("Usage:\n  /load <modelname>")
-				continue
-			}
-			opts.Model = args[1]
-			opts.Messages = []api.Message{}
-			fmt.Printf("Loading model '%s'\n", opts.Model)
-			if err := loadModel(cmd, &opts); err != nil {
-				return err
-			}
-			continue
-		case strings.HasPrefix(line, "/save"):
-			args := strings.Fields(line)
-			if len(args) != 2 {
-				fmt.Println("Usage:\n  /save <modelname>")
-				continue
-			}
-
-			client, err := api.ClientFromEnvironment()
-			if err != nil {
-				fmt.Println("error: couldn't connect to ollama server")
-				return err
-			}
-
-			req := &api.CreateRequest{
-				Name:      args[1],
-				Modelfile: buildModelfile(opts),
-			}
-			fn := func(resp api.ProgressResponse) error { return nil }
-			err = client.Create(cmd.Context(), req, fn)
-			if err != nil {
-				if strings.Contains(err.Error(), errtypes.InvalidModelNameErrMsg) {
-					fmt.Printf("error: The model name '%s' is invalid\n", args[1])
-					continue
-				}
-				return err
-			}
-			fmt.Printf("Created new model '%s'\n", args[1])
-			continue
-		case strings.HasPrefix(line, "/clear"):
-			opts.Messages = []api.Message{}
-			if opts.System != "" {
-				newMessage := api.Message{Role: "system", Content: opts.System}
-				opts.Messages = append(opts.Messages, newMessage)
-			}
-			fmt.Println("Cleared session context")
-			continue
-		case strings.HasPrefix(line, "/set"):
-			args := strings.Fields(line)
-			if len(args) > 1 {
-				switch args[1] {
-				case "history":
-					scanner.HistoryEnable()
-				case "nohistory":
-					scanner.HistoryDisable()
-				case "wordwrap":
-					opts.WordWrap = true
-					fmt.Println("Set 'wordwrap' mode.")
-				case "nowordwrap":
-					opts.WordWrap = false
-					fmt.Println("Set 'nowordwrap' mode.")
-				case "verbose":
-					if err := cmd.Flags().Set("verbose", "true"); err != nil {
-						return err
-					}
-					fmt.Println("Set 'verbose' mode.")
-				case "quiet":
-					if err := cmd.Flags().Set("verbose", "false"); err != nil {
-						return err
-					}
-					fmt.Println("Set 'quiet' mode.")
-				case "format":
-					if len(args) < 3 || args[2] != "json" {
-						fmt.Println("Invalid or missing format. For 'json' mode use '/set format json'")
-					} else {
-						opts.Format = args[2]
-						fmt.Printf("Set format to '%s' mode.\n", args[2])
-					}
-				case "noformat":
-					opts.Format = ""
-					fmt.Println("Disabled format.")
-				case "parameter":
-					if len(args) < 4 {
-						usageParameters()
-						continue
-					}
-					params := args[3:]
-					fp, err := api.FormatParams(map[string][]string{args[2]: params})
-					if err != nil {
-						fmt.Printf("Couldn't set parameter: %q\n", err)
-						continue
-					}
-					fmt.Printf("Set parameter '%s' to '%s'\n", args[2], strings.Join(params, ", "))
-					opts.Options[args[2]] = fp[args[2]]
-				case "system":
-					if len(args) < 3 {
-						usageSet()
-						continue
-					}
-
-					multiline = MultilineSystem
-
-					line := strings.Join(args[2:], " ")
-					line, ok := strings.CutPrefix(line, `"""`)
-					if !ok {
-						multiline = MultilineNone
-					} else {
-						// only cut suffix if the line is multiline
-						line, ok = strings.CutSuffix(line, `"""`)
-						if ok {
-							multiline = MultilineNone
-						}
-					}
-
-					sb.WriteString(line)
-					if multiline != MultilineNone {
-						scanner.Prompt.UseAlt = true
-						continue
-					}
-
-					opts.System = sb.String() // for display in modelfile
-					newMessage := api.Message{Role: "system", Content: sb.String()}
-					// Check if the slice is not empty and the last message is from 'system'
-					if len(opts.Messages) > 0 && opts.Messages[len(opts.Messages)-1].Role == "system" {
-						// Replace the last message
-						opts.Messages[len(opts.Messages)-1] = newMessage
-					} else {
-						opts.Messages = append(opts.Messages, newMessage)
-					}
-					fmt.Println("Set system message.")
-					sb.Reset()
-
-					sb.Reset()
-					continue
-				default:
-					fmt.Printf("Unknown command '/set %s'. Type /? for help\n", args[1])
-				}
-			} else {
-				usageSet()
-			}
-		case strings.HasPrefix(line, "/show"):
-			args := strings.Fields(line)
-			if len(args) > 1 {
-				client, err := api.ClientFromEnvironment()
-				if err != nil {
-					fmt.Println("error: couldn't connect to ollama server")
-					return err
-				}
-				req := &api.ShowRequest{
-					Name:    opts.Model,
-					System:  opts.System,
-					Options: opts.Options,
-				}
-				resp, err := client.Show(cmd.Context(), req)
-				if err != nil {
-					fmt.Println("error: couldn't get model")
-					return err
-				}
-
-				switch args[1] {
-				case "info":
-					showInfo(resp)
-				case "license":
-					if resp.License == "" {
-						fmt.Println("No license was specified for this model.")
-					} else {
-						fmt.Println(resp.License)
-					}
-				case "modelfile":
-					fmt.Println(resp.Modelfile)
-				case "parameters":
-					if resp.Parameters == "" {
-						fmt.Println("No parameters were specified for this model.")
-					} else {
-						if len(opts.Options) > 0 {
-							fmt.Println("User defined parameters:")
-							for k, v := range opts.Options {
-								fmt.Printf("%-*s %v\n", 30, k, v)
-							}
-							fmt.Println()
-						}
-						fmt.Println("Model defined parameters:")
-						fmt.Println(resp.Parameters)
-					}
-				case "system":
-					switch {
-					case opts.System != "":
-						fmt.Println(opts.System + "\n")
-					case resp.System != "":
-						fmt.Println(resp.System + "\n")
-					default:
-						fmt.Println("No system message was specified for this model.")
-					}
-				case "template":
-					if resp.Template != "" {
-						fmt.Println(resp.Template)
-					} else {
-						fmt.Println("No prompt template was specified for this model.")
-					}
-				default:
-					fmt.Printf("Unknown command '/show %s'. Type /? for help\n", args[1])
-				}
-			} else {
-				usageShow()
-			}
-		case strings.HasPrefix(line, "/help"), strings.HasPrefix(line, "/?"):
-			args := strings.Fields(line)
-			if len(args) > 1 {
-				switch args[1] {
-				case "set", "/set":
-					usageSet()
-				case "show", "/show":
-					usageShow()
-				case "shortcut", "shortcuts":
-					usageShortcuts()
-				}
-			} else {
-				usage()
-			}
-		case strings.HasPrefix(line, "/exit"), strings.HasPrefix(line, "/bye"):
-			return nil
-		case strings.HasPrefix(line, "/"):
-			args := strings.Fields(line)
-			isFile := false
-
-			if opts.MultiModal {
-				for _, f := range extractFileNames(line) {
-					if strings.HasPrefix(f, args[0]) {
-						isFile = true
-						break
-					}
-				}
-			}
-
-			if !isFile {
-				fmt.Printf("Unknown command '%s'. Type /? for help\n", args[0])
-				continue
-			}
-
-			sb.WriteString(line)
-		default:
-			sb.WriteString(line)
-		}
-
-		if sb.Len() > 0 && multiline == MultilineNone {
-			newMessage := api.Message{Role: "user", Content: sb.String()}
-
-			if opts.MultiModal {
-				msg, images, err := extractFileData(sb.String())
-				if err != nil {
-					return err
-				}
-
-				// clear all previous images for better responses
-				if len(images) > 0 {
-					for i := range opts.Messages {
-						opts.Messages[i].Images = nil
-					}
-				}
-
-				newMessage.Content = msg
-				newMessage.Images = images
-			}
-
-			opts.Messages = append(opts.Messages, newMessage)
-
-			assistant, err := chat(cmd, opts)
-			if err != nil {
-				return err
-			}
-			if assistant != nil {
-				opts.Messages = append(opts.Messages, *assistant)
-			}
-
-			sb.Reset()
-		}
-	}
-}
-
-func buildModelfile(opts runOptions) string {
-	var f parser.File
-	f.Commands = append(f.Commands, parser.Command{Name: "model", Args: cmp.Or(opts.ParentModel, opts.Model)})
-
-	if opts.System != "" {
-		f.Commands = append(f.Commands, parser.Command{Name: "system", Args: opts.System})
-	}
-
-	keys := maps.Keys(opts.Options)
-	slices.Sort(keys)
-	for _, k := range keys {
-		v := opts.Options[k]
-		var cmds []parser.Command
-		switch t := v.(type) {
-		case []string:
-			for _, s := range t {
-				cmds = append(cmds, parser.Command{Name: k, Args: s})
-			}
-		default:
-			cmds = append(cmds, parser.Command{Name: k, Args: fmt.Sprintf("%v", t)})
-		}
-
-		f.Commands = append(f.Commands, cmds...)
-	}
-
-	for _, msg := range opts.Messages {
-		f.Commands = append(f.Commands, parser.Command{Name: "message", Args: fmt.Sprintf("%s: %s", msg.Role, msg.Content)})
-	}
-
-	return f.String()
-}
-
-func normalizeFilePath(fp string) string {
-	// Define a map of escaped characters and their replacements
-	replacements := map[string]string{
-		"\\ ":  " ",  // Escaped space
-		"\\(":  "(",  // Escaped left parenthesis
-		"\\)":  ")",  // Escaped right parenthesis
-		"\\[":  "[",  // Escaped left square bracket
-		"\\]":  "]",  // Escaped right square bracket
-		"\\{":  "{",  // Escaped left curly brace
-		"\\}":  "}",  // Escaped right curly brace
-		"\\$":  "$",  // Escaped dollar sign
-		"\\&":  "&",  // Escaped ampersand
-		"\\;":  ";",  // Escaped semicolon
-		"\\'":  "'",  // Escaped single quote
-		"\\\\": "\\", // Escaped backslash
-		"\\*":  "*",  // Escaped asterisk
-		"\\?":  "?",  // Escaped question mark
-	}
-
-	for escaped, actual := range replacements {
-		fp = strings.ReplaceAll(fp, escaped, actual)
-	}
-	return fp
-}
-
-func extractFileNames(input string) []string {
-	// Regex to match file paths starting with optional drive letter, / ./ \ or .\ and include escaped or unescaped spaces (\ or %20)
-	// and followed by more characters and a file extension
-	// This will capture non filename strings, but we'll check for file existence to remove mismatches
-	regexPattern := `(?:[a-zA-Z]:)?(?:\./|/|\\)[\S\\ ]+?\.(?i:jpg|jpeg|png|svg)\b`
-	re := regexp.MustCompile(regexPattern)
-
-	return re.FindAllString(input, -1)
-}
-
-func extractFileData(input string) (string, []api.ImageData, error) {
-	filePaths := extractFileNames(input)
-	var imgs []api.ImageData
-
-	for _, fp := range filePaths {
-		nfp := normalizeFilePath(fp)
-		data, err := getImageData(nfp)
-		if err != nil {
-			if os.IsNotExist(err) {
-				continue
-			}
-			fmt.Fprintf(os.Stderr, "Couldn't process image: %q\n", err)
-			return "", imgs, err
-		}
-		fmt.Fprintf(os.Stderr, "Added image '%s'\n", nfp)
-		input = strings.ReplaceAll(input, fp, "")
-		imgs = append(imgs, data)
-	}
-	return input, imgs, nil
-}
-
-func getImageData(filePath string) ([]byte, error) {
-	file, err := os.Open(filePath)
-	if err != nil {
-		return nil, err
-	}
-	defer file.Close()
-
-	buf := make([]byte, 512)
-	_, err = file.Read(buf)
-	if err != nil {
-		return nil, err
-	}
-
-	contentType := http.DetectContentType(buf)
-	allowedTypes := []string{"image/jpeg", "image/jpg", "image/png"}
-	if !slices.Contains(allowedTypes, contentType) {
-		return nil, fmt.Errorf("invalid image type: %s", contentType)
-	}
-
-	info, err := file.Stat()
-	if err != nil {
-		return nil, err
-	}
-
-	// Check if the file size exceeds 100MB
-	var maxSize int64 = 100 * 1024 * 1024 // 100MB in bytes
-	if info.Size() > maxSize {
-		return nil, fmt.Errorf("file size exceeds maximum limit (100MB)")
-	}
-
-	buf = make([]byte, info.Size())
-	_, err = file.Seek(0, 0)
-	if err != nil {
-		return nil, err
-	}
-
-	_, err = io.ReadFull(file, buf)
-	if err != nil {
-		return nil, err
-	}
-
-	return buf, nil
-}

cmd/interactive_test.go

@@ -1,107 +0,0 @@
-package cmd
-
-import (
-	"testing"
-
-	"github.com/google/go-cmp/cmp"
-	"github.com/stretchr/testify/assert"
-
-	"github.com/ollama/ollama/api"
-)
-
-func TestExtractFilenames(t *testing.T) {
-	// Unix style paths
-	input := ` some preamble 
- ./relative\ path/one.png inbetween1 ./not a valid two.jpg inbetween2
-/unescaped space /three.jpeg inbetween3 /valid\ path/dir/four.png "./quoted with spaces/five.svg`
-	res := extractFileNames(input)
-	assert.Len(t, res, 5)
-	assert.Contains(t, res[0], "one.png")
-	assert.Contains(t, res[1], "two.jpg")
-	assert.Contains(t, res[2], "three.jpeg")
-	assert.Contains(t, res[3], "four.png")
-	assert.Contains(t, res[4], "five.svg")
-	assert.NotContains(t, res[4], '"')
-	assert.NotContains(t, res, "inbtween")
-
-	// Windows style paths
-	input = ` some preamble
- c:/users/jdoe/one.png inbetween1 c:/program files/someplace/two.jpg inbetween2 
- /absolute/nospace/three.jpeg inbetween3 /absolute/with space/four.png inbetween4
-./relative\ path/five.svg inbetween5 "./relative with/spaces/six.png inbetween6
-d:\path with\spaces\seven.svg inbetween7 c:\users\jdoe\eight.png inbetween8 
- d:\program files\someplace\nine.png inbetween9 "E:\program files\someplace\ten.svg some ending
-`
-	res = extractFileNames(input)
-	assert.Len(t, res, 10)
-	assert.NotContains(t, res, "inbtween")
-	assert.Contains(t, res[0], "one.png")
-	assert.Contains(t, res[0], "c:")
-	assert.Contains(t, res[1], "two.jpg")
-	assert.Contains(t, res[1], "c:")
-	assert.Contains(t, res[2], "three.jpeg")
-	assert.Contains(t, res[3], "four.png")
-	assert.Contains(t, res[4], "five.svg")
-	assert.Contains(t, res[5], "six.png")
-	assert.Contains(t, res[6], "seven.svg")
-	assert.Contains(t, res[6], "d:")
-	assert.Contains(t, res[7], "eight.png")
-	assert.Contains(t, res[7], "c:")
-	assert.Contains(t, res[8], "nine.png")
-	assert.Contains(t, res[8], "d:")
-	assert.Contains(t, res[9], "ten.svg")
-	assert.Contains(t, res[9], "E:")
-}
-
-func TestModelfileBuilder(t *testing.T) {
-	opts := runOptions{
-		Model:  "hork",
-		System: "You are part horse and part shark, but all hork. Do horklike things",
-		Messages: []api.Message{
-			{Role: "user", Content: "Hey there hork!"},
-			{Role: "assistant", Content: "Yes it is true, I am half horse, half shark."},
-		},
-		Options: map[string]any{
-			"temperature":      0.9,
-			"seed":             42,
-			"penalize_newline": false,
-			"stop":             []string{"hi", "there"},
-		},
-	}
-
-	t.Run("model", func(t *testing.T) {
-		expect := `FROM hork
-SYSTEM You are part horse and part shark, but all hork. Do horklike things
-PARAMETER penalize_newline false
-PARAMETER seed 42
-PARAMETER stop hi
-PARAMETER stop there
-PARAMETER temperature 0.9
-MESSAGE user Hey there hork!
-MESSAGE assistant Yes it is true, I am half horse, half shark.
-`
-
-		actual := buildModelfile(opts)
-		if diff := cmp.Diff(expect, actual); diff != "" {
-			t.Errorf("mismatch (-want +got):\n%s", diff)
-		}
-	})
-
-	t.Run("parent model", func(t *testing.T) {
-		opts.ParentModel = "horseshark"
-		expect := `FROM horseshark
-SYSTEM You are part horse and part shark, but all hork. Do horklike things
-PARAMETER penalize_newline false
-PARAMETER seed 42
-PARAMETER stop hi
-PARAMETER stop there
-PARAMETER temperature 0.9
-MESSAGE user Hey there hork!
-MESSAGE assistant Yes it is true, I am half horse, half shark.
-`
-		actual := buildModelfile(opts)
-		if diff := cmp.Diff(expect, actual); diff != "" {
-			t.Errorf("mismatch (-want +got):\n%s", diff)
-		}
-	})
-}

cmd/spinner.go

@@ -0,0 +1,44 @@
+package cmd
+
+import (
+	"fmt"
+	"os"
+	"time"
+
+	"github.com/jmorganca/ollama/progressbar"
+)
+
+type Spinner struct {
+	description string
+	*progressbar.ProgressBar
+}
+
+func NewSpinner(description string) *Spinner {
+	return &Spinner{
+		description: description,
+		ProgressBar: progressbar.NewOptions(-1,
+			progressbar.OptionSetWriter(os.Stderr),
+			progressbar.OptionThrottle(60*time.Millisecond),
+			progressbar.OptionSpinnerType(14),
+			progressbar.OptionSetRenderBlankState(true),
+			progressbar.OptionSetElapsedTime(false),
+			progressbar.OptionClearOnFinish(),
+			progressbar.OptionSetDescription(description),
+		),
+	}
+}
+
+func (s *Spinner) Spin(tick time.Duration) {
+	for range time.Tick(tick) {
+		if s.IsFinished() {
+			break
+		}
+
+		s.Add(1)
+	}
+}
+
+func (s *Spinner) Stop() {
+	s.Finish()
+	fmt.Println(s.description)
+}

cmd/start.go

@@ -1,27 +0,0 @@
-//go:build darwin || windows
-
-package cmd
-
-import (
-	"context"
-	"errors"
-	"time"
-
-	"github.com/ollama/ollama/api"
-)
-
-func waitForServer(ctx context.Context, client *api.Client) error {
-	// wait for the server to start
-	timeout := time.After(5 * time.Second)
-	tick := time.Tick(500 * time.Millisecond)
-	for {
-		select {
-		case <-timeout:
-			return errors.New("timed out waiting for server to start")
-		case <-tick:
-			if err := client.Heartbeat(ctx); err == nil {
-				return nil // server has started
-			}
-		}
-	}
-}

cmd/start_darwin.go

@@ -1,30 +0,0 @@
-package cmd
-
-import (
-	"context"
-	"fmt"
-	"os"
-	"os/exec"
-	"strings"
-
-	"github.com/ollama/ollama/api"
-)
-
-func startApp(ctx context.Context, client *api.Client) error {
-	exe, err := os.Executable()
-	if err != nil {
-		return err
-	}
-	link, err := os.Readlink(exe)
-	if err != nil {
-		return err
-	}
-	if !strings.Contains(link, "Ollama.app") {
-		return fmt.Errorf("could not find ollama app")
-	}
-	path := strings.Split(link, "Ollama.app")
-	if err := exec.Command("/usr/bin/open", "-a", path[0]+"Ollama.app").Run(); err != nil {
-		return err
-	}
-	return waitForServer(ctx, client)
-}

cmd/start_default.go

@@ -1,14 +0,0 @@
-//go:build !windows && !darwin
-
-package cmd
-
-import (
-	"context"
-	"fmt"
-
-	"github.com/ollama/ollama/api"
-)
-
-func startApp(ctx context.Context, client *api.Client) error {
-	return fmt.Errorf("could not connect to ollama server, run 'ollama serve' to start it")
-}

cmd/start_windows.go

@@ -1,58 +0,0 @@
-package cmd
-
-import (
-	"context"
-	"errors"
-	"fmt"
-	"os"
-	"os/exec"
-	"path/filepath"
-	"strings"
-	"syscall"
-
-	"github.com/ollama/ollama/api"
-)
-
-func startApp(ctx context.Context, client *api.Client) error {
-	// log.Printf("XXX Attempting to find and start ollama app")
-	AppName := "ollama app.exe"
-	exe, err := os.Executable()
-	if err != nil {
-		return err
-	}
-	appExe := filepath.Join(filepath.Dir(exe), AppName)
-	_, err = os.Stat(appExe)
-	if errors.Is(err, os.ErrNotExist) {
-		// Try the standard install location
-		localAppData := os.Getenv("LOCALAPPDATA")
-		appExe = filepath.Join(localAppData, "Ollama", AppName)
-		_, err := os.Stat(appExe)
-		if errors.Is(err, os.ErrNotExist) {
-			// Finally look in the path
-			appExe, err = exec.LookPath(AppName)
-			if err != nil {
-				return fmt.Errorf("could not locate ollama app")
-			}
-		}
-	}
-	// log.Printf("XXX attempting to start app %s", appExe)
-
-	cmd_path := "c:\\Windows\\system32\\cmd.exe"
-	cmd := exec.Command(cmd_path, "/c", appExe)
-	// TODO - these hide flags aren't working - still pops up a command window for some reason
-	cmd.SysProcAttr = &syscall.SysProcAttr{CreationFlags: 0x08000000, HideWindow: true}
-
-	// TODO this didn't help either...
-	cmd.Stdin = strings.NewReader("")
-	cmd.Stdout = os.Stdout
-	cmd.Stderr = os.Stderr
-
-	if err := cmd.Start(); err != nil {
-		return fmt.Errorf("unable to start ollama app %w", err)
-	}
-
-	if cmd.Process != nil {
-		defer cmd.Process.Release() //nolint:errcheck
-	}
-	return waitForServer(ctx, client)
-}

convert/convert.go

@@ -1,122 +0,0 @@
-package convert
-
-import (
-	"encoding/json"
-	"errors"
-	"fmt"
-	"io"
-	"io/fs"
-	"log/slog"
-
-	"github.com/ollama/ollama/llm"
-)
-
-type Parameters struct {
-	Architectures []string `json:"architectures"`
-	VocabSize     uint32   `json:"vocab_size"`
-}
-
-func (Parameters) KV(t *Tokenizer) llm.KV {
-	kv := llm.KV{
-		"general.file_type":            uint32(1),
-		"general.quantization_version": uint32(2),
-		"tokenizer.ggml.pre":           t.Pre,
-		"tokenizer.ggml.model":         t.Vocabulary.Model,
-		"tokenizer.ggml.tokens":        t.Vocabulary.Tokens,
-		"tokenizer.ggml.scores":        t.Vocabulary.Scores,
-		"tokenizer.ggml.token_type":    t.Vocabulary.Types,
-	}
-
-	if t.Template != "" {
-		kv["tokenizer.chat_template"] = t.Template
-	}
-
-	for _, sv := range t.SpecialVocabulary {
-		kv[fmt.Sprintf("tokenizer.ggml.%s_token_id", sv.Key())] = uint32(sv.ID)
-		kv[fmt.Sprintf("tokenizer.ggml.add_%s_token", sv.Key())] = sv.AddToken
-	}
-
-	return kv
-}
-
-func (Parameters) specialTokenTypes() []string {
-	return []string{
-		"bos", "eos", "unk", "sep", "pad", "cls", "mask",
-	}
-}
-
-func (Parameters) writeFile(ws io.WriteSeeker, kv llm.KV, ts []llm.Tensor) error {
-	return llm.WriteGGUF(ws, kv, ts)
-}
-
-type Converter interface {
-	// KV maps parameters to LLM key-values
-	KV(*Tokenizer) llm.KV
-	// Tensors maps input tensors to LLM tensors. Model specific modifications can be done here.
-	Tensors([]Tensor) []llm.Tensor
-
-	// tensorName returns the LLM tensor name for a specific input name
-	tensorName(string) string
-	// specialTokenTypes returns any special token types the model uses
-	specialTokenTypes() []string
-	writeFile(io.WriteSeeker, llm.KV, []llm.Tensor) error
-}
-
-// Convert writes an Ollama compatible model to the provided io.WriteSeeker based on configurations
-// and files it finds in the input path.
-// Supported input model formats include safetensors.
-// Supported input tokenizers files include tokenizer.json (preferred) and tokenizer.model.
-func Convert(fsys fs.FS, ws io.WriteSeeker) error {
-	bts, err := fs.ReadFile(fsys, "config.json")
-	if err != nil {
-		return err
-	}
-
-	var p Parameters
-	if err := json.Unmarshal(bts, &p); err != nil {
-		return err
-	}
-
-	if len(p.Architectures) < 1 {
-		return errors.New("unknown architecture")
-	}
-
-	var conv Converter
-	switch p.Architectures[0] {
-	case "LlamaForCausalLM", "MistralForCausalLM":
-		conv = &llama{}
-	case "MixtralForCausalLM":
-		conv = &mixtral{}
-	case "GemmaForCausalLM":
-		conv = &gemma{}
-	default:
-		return errors.New("unsupported architecture")
-	}
-
-	if err := json.Unmarshal(bts, conv); err != nil {
-		return err
-	}
-
-	t, err := parseTokenizer(fsys, conv.specialTokenTypes())
-	if err != nil {
-		return err
-	}
-
-	if vocabSize := int(p.VocabSize); vocabSize > len(t.Vocabulary.Tokens) {
-		slog.Warn("vocabulary is smaller than expected, padding with dummy tokens", "expect", p.VocabSize, "actual", len(t.Vocabulary.Tokens))
-		for i := range vocabSize - len(t.Vocabulary.Tokens) {
-			t.Vocabulary.Tokens = append(t.Vocabulary.Tokens, fmt.Sprintf("[PAD%d]", i))
-			t.Vocabulary.Scores = append(t.Vocabulary.Scores, -1)
-			t.Vocabulary.Types = append(t.Vocabulary.Types, tokenTypeUserDefined)
-		}
-	} else {
-		slog.Debug("vocabulary", "size", len(t.Vocabulary.Tokens))
-	}
-
-	ts, err := parseTensors(fsys)
-	if err != nil {
-		return err
-	}
-
-	return conv.writeFile(ws, conv.KV(t), conv.Tensors(ts))
-}

convert/convert_gemma.go

@@ -1,103 +0,0 @@
-package convert
-
-import (
-	"strings"
-
-	"github.com/pdevine/tensor"
-	"github.com/pdevine/tensor/native"
-
-	"github.com/ollama/ollama/llm"
-)
-
-type gemma struct {
-	Parameters
-	MaxPositionEmbeddings uint32  `json:"max_position_embeddings"`
-	HiddenSize            uint32  `json:"hidden_size"`
-	HiddenLayers          uint32  `json:"num_hidden_layers"`
-	IntermediateSize      uint32  `json:"intermediate_size"`
-	NumAttentionHeads     uint32  `json:"num_attention_heads"`
-	NumKeyValueHeads      uint32  `json:"num_key_value_heads"`
-	RMSNormEPS            float32 `json:"rms_norm_eps"`
-	HeadDim               uint32  `json:"head_dim"`
-}
-
-var _ Converter = (*gemma)(nil)
-
-func (p *gemma) KV(t *Tokenizer) llm.KV {
-	kv := p.Parameters.KV(t)
-	kv["general.architecture"] = "gemma"
-	kv["general.name"] = "gemma"
-	kv["gemma.context_length"] = p.MaxPositionEmbeddings
-	kv["gemma.embedding_length"] = p.HiddenSize
-	kv["gemma.block_count"] = p.HiddenLayers
-	kv["gemma.feed_forward_length"] = p.IntermediateSize
-	kv["gemma.attention.head_count"] = p.NumAttentionHeads
-	kv["gemma.attention.head_count_kv"] = p.NumKeyValueHeads
-	kv["gemma.attention.layer_norm_rms_epsilon"] = p.RMSNormEPS
-	kv["gemma.attention.key_length"] = p.HeadDim
-	kv["gemma.attention.value_length"] = p.HeadDim
-	kv["tokenizer.ggml.eot_token_id"] = uint32(107)
-	kv["tokenizer.ggml.middle_token_id"] = uint32(68)
-	kv["tokenizer.ggml.prefix_token_id"] = uint32(67)
-	kv["tokenizer.ggml.suffix_token_id"] = uint32(69)
-	return kv
-}
-
-func (p *gemma) Tensors(ts []Tensor) []llm.Tensor {
-	var out []llm.Tensor
-	for _, t := range ts {
-		name := p.tensorName(t.Name())
-		if strings.HasSuffix(name, "_norm.weight") {
-			t.SetRepacker(p.addOne)
-		}
-
-		out = append(out, llm.Tensor{
-			Name:     name,
-			Kind:     t.Kind(),
-			Shape:    t.Shape(),
-			WriterTo: t,
-		})
-	}
-
-	return out
-}
-
-func (p *gemma) tensorName(n string) string {
-	return strings.NewReplacer(
-		"model.embed_tokens", "token_embd",
-		"model.norm", "output_norm",
-		"model.layers", "blk",
-		"input_layernorm", "attn_norm",
-		"self_attn.q_proj", "attn_q",
-		"self_attn.k_proj", "attn_k",
-		"self_attn.v_proj", "attn_v",
-		"self_attn.o_proj", "attn_output",
-		"mlp.gate_proj", "ffn_gate",
-		"mlp.down_proj", "ffn_down",
-		"mlp.up_proj", "ffn_up",
-		"post_attention_layernorm", "ffn_norm",
-		"block_sparse_moe.gate", "ffn_inp",
-	).Replace(n)
-}
-
-func (*gemma) addOne(_ string, data []float32, shape []uint64) ([]float32, error) {
-	n := tensor.New(tensor.WithShape(int(shape[0])), tensor.WithBacking(data))
-	ones := tensor.Ones(tensor.Float32, int(shape[0]))
-
-	n, err := n.Add(ones)
-	if err != nil {
-		return nil, err
-	}
-
-	ts, err := native.SelectF32(n, 0)
-	if err != nil {
-		return nil, err
-	}
-
-	var f32s []float32
-	for _, t := range ts {
-		f32s = append(f32s, t...)
-	}
-
-	return f32s, nil
-}

convert/convert_llama.go

@@ -1,182 +0,0 @@
-package convert
-
-import (
-	"cmp"
-	"fmt"
-	"strings"
-
-	"github.com/ollama/ollama/llm"
-	"github.com/pdevine/tensor"
-	"github.com/pdevine/tensor/native"
-)
-
-type llama struct {
-	Parameters
-	NLayers               uint32  `json:"n_layers"`
-	NumHiddenLayers       uint32  `json:"num_hidden_layers"`
-	NLayer                uint32  `json:"n_layer"`
-	MaxPositionEmbeddings uint32  `json:"max_position_embeddings"`
-	NCtx                  uint32  `json:"n_ctx"`
-	HiddenSize            uint32  `json:"hidden_size"`
-	NEmbd                 uint32  `json:"n_embd"`
-	IntermediateSize      uint32  `json:"intermediate_size"`
-	NInner                uint32  `json:"n_inner"`
-	NumAttentionHeads     uint32  `json:"num_attention_heads"`
-	NHead                 uint32  `json:"n_head"`
-	NumKeyValueHeads      uint32  `json:"num_key_value_heads"`
-	RopeTheta             float32 `json:"rope_theta"`
-	RopeScaling           struct {
-		Type   string  `json:"type"`
-		Factor float32 `json:"factor"`
-	} `json:"rope_scaling"`
-	RMSNormEPS       float32 `json:"rms_norm_eps"`
-	LayerNormEPS     float32 `json:"layer_norm_eps"`
-	LayerNormEpsilon float32 `json:"layer_norm_epsilon"`
-	NormEpsilon      float32 `json:"norm_epsilon"`
-	HeadDim          uint32  `json:"head_dim"`
-}
-
-var _ Converter = (*llama)(nil)
-
-func (p *llama) KV(t *Tokenizer) llm.KV {
-	kv := p.Parameters.KV(t)
-	kv["general.architecture"] = "llama"
-	kv["general.name"] = "llama"
-	kv["llama.vocab_size"] = p.VocabSize
-
-	kv["llama.block_count"] = cmp.Or(p.NLayers, p.NumHiddenLayers, p.NLayer)
-
-	if contextLength := cmp.Or(p.MaxPositionEmbeddings, p.NCtx); contextLength > 0 {
-		kv["llama.context_length"] = contextLength
-	}
-
-	if embeddingLength := cmp.Or(p.HiddenSize, p.NEmbd); embeddingLength > 0 {
-		kv["llama.embedding_length"] = cmp.Or(p.HiddenSize, p.NEmbd)
-	}
-
-	if feedForwardLength := cmp.Or(p.IntermediateSize, p.NInner); feedForwardLength > 0 {
-		kv["llama.feed_forward_length"] = cmp.Or(p.IntermediateSize, p.NInner)
-	}
-
-	if headCount := cmp.Or(p.NumAttentionHeads, p.NHead); headCount > 0 {
-		kv["llama.attention.head_count"] = cmp.Or(p.NumAttentionHeads, p.NHead)
-		kv["llama.rope.dimension_count"] = p.HiddenSize / headCount
-	}
-
-	if p.RopeTheta > 0 {
-		kv["llama.rope.freq_base"] = p.RopeTheta
-	}
-
-	if p.RopeScaling.Type == "linear" {
-		kv["llama.rope.scaling.type"] = p.RopeScaling.Type
-		kv["llama.rope.scaling.factor"] = p.RopeScaling.Factor
-	}
-
-	if p.NumKeyValueHeads > 0 {
-		kv["llama.attention.head_count_kv"] = p.NumKeyValueHeads
-	}
-
-	if p.RMSNormEPS > 0 {
-		kv["llama.attention.layer_norm_rms_epsilon"] = p.RMSNormEPS
-	}
-
-	if layerNormEpsilon := cmp.Or(p.LayerNormEPS, p.LayerNormEpsilon, p.NormEpsilon); layerNormEpsilon > 0 {
-		kv["llama.attention.layer_norm_epsilon"] = layerNormEpsilon
-	}
-
-	if p.HeadDim > 0 {
-		kv["llama.attention.key_length"] = p.HeadDim
-		kv["llama.attention.value_length"] = p.HeadDim
-	}
-
-	if len(t.Merges) > 0 {
-		kv["tokenizer.ggml.merges"] = t.Merges
-	}
-
-	return kv
-}
-
-func (p *llama) Tensors(ts []Tensor) []llm.Tensor {
-	var out []llm.Tensor
-	for _, t := range ts {
-		name := p.tensorName(t.Name())
-		if strings.HasSuffix(name, "attn_q.weight") ||
-			strings.HasSuffix(name, "attn_k.weight") {
-			t.SetRepacker(p.repack)
-		}
-
-		out = append(out, llm.Tensor{
-			Name:     name,
-			Kind:     t.Kind(),
-			Shape:    t.Shape(),
-			WriterTo: t,
-		})
-	}
-
-	return out
-}
-
-func (p *llama) tensorName(n string) string {
-	return strings.NewReplacer(
-		"lm_head", "output",
-		"model.embed_tokens", "token_embd",
-		"model.norm", "output_norm",
-		"model.layers", "blk",
-		"input_layernorm", "attn_norm",
-		"self_attn.q_proj", "attn_q",
-		"self_attn.k_proj", "attn_k",
-		"self_attn.v_proj", "attn_v",
-		"self_attn.o_proj", "attn_output",
-		"mlp.gate_proj", "ffn_gate",
-		"mlp.down_proj", "ffn_down",
-		"mlp.up_proj", "ffn_up",
-		"post_attention_layernorm", "ffn_norm",
-		// mixtral
-		"block_sparse_moe.gate", "ffn_gate_inp",
-	).Replace(n)
-}
-
-func (p *llama) repack(name string, data []float32, shape []uint64) ([]float32, error) {
-	var dims []int
-	for _, dim := range shape {
-		dims = append(dims, int(dim))
-	}
-
-	var heads uint32
-	if strings.HasSuffix(name, "q_proj.weight") {
-		heads = p.NumAttentionHeads
-	} else if strings.HasSuffix(name, "k_proj.weight") {
-		heads = cmp.Or(p.NumKeyValueHeads, p.NumAttentionHeads)
-	} else {
-		return nil, fmt.Errorf("unknown tensor for repack: %s", name)
-	}
-
-	n := tensor.New(tensor.WithShape(dims...), tensor.WithBacking(data))
-	if err := n.Reshape(append([]int{int(heads), 2, dims[0] / int(heads) / 2}, dims[1:]...)...); err != nil {
-		return nil, err
-	}
-
-	if err := n.T(0, 2, 1, 3); err != nil {
-		return nil, err
-	}
-
-	if err := n.Reshape(dims...); err != nil {
-		return nil, err
-	}
-
-	if err := n.Transpose(); err != nil {
-		return nil, err
-	}
-
-	ts, err := native.SelectF32(n, 1)
-	if err != nil {
-		return nil, err
-	}
-
-	var f32s []float32
-	for _, t := range ts {
-		f32s = append(f32s, t...)
-	}
-
-	return f32s, nil
-}

convert/convert_mixtral.go

@@ -1,89 +0,0 @@
-package convert
-
-import (
-	"fmt"
-	"io"
-	"slices"
-	"strings"
-
-	"github.com/ollama/ollama/llm"
-)
-
-type mixtral struct {
-	llama
-	NumLocalExperts    uint32 `json:"num_local_experts"`
-	NumExpertsPerToken uint32 `json:"num_experts_per_tok"`
-}
-
-var _ Converter = (*mixtral)(nil)
-
-func (p *mixtral) KV(t *Tokenizer) llm.KV {
-	kv := p.llama.KV(t)
-
-	if p.NumLocalExperts > 0 {
-		kv["llama.expert_count"] = p.NumLocalExperts
-	}
-
-	if p.NumExpertsPerToken > 0 {
-		kv["llama.expert_used_count"] = p.NumExpertsPerToken
-	}
-
-	return kv
-}
-
-func (p *mixtral) Tensors(ts []Tensor) []llm.Tensor {
-	oldnew := []string{
-		"model.layers", "blk",
-		"w1", "ffn_gate_exps",
-		"w2", "ffn_down_exps",
-		"w3", "ffn_up_exps",
-	}
-
-	for i := range p.NumLocalExperts {
-		oldnew = append(oldnew, fmt.Sprintf(".block_sparse_moe.experts.%d.", i), ".")
-	}
-
-	// group experts of the same layer (model.layers.%d) and type (w[123]) into a single tensor
-	namer := strings.NewReplacer(oldnew...)
-	experts := make(map[string]experts)
-
-	// merge experts into a single tensor while removing them from ts
-	ts = slices.DeleteFunc(ts, func(t Tensor) bool {
-		if !strings.Contains(t.Name(), ".block_sparse_moe.experts.") {
-			return false
-		}
-
-		name := namer.Replace(t.Name())
-		experts[name] = append(experts[name], t)
-		return true
-	})
-
-	var out []llm.Tensor
-	for n, e := range experts {
-		// TODO(mxyng): sanity check experts
-		out = append(out, llm.Tensor{
-			Name:     n,
-			Kind:     e[0].Kind(),
-			Shape:    append([]uint64{uint64(len(e))}, e[0].Shape()...),
-			WriterTo: e,
-		})
-	}
-
-	return append(out, p.llama.Tensors(ts)...)
-}
-
-type experts []Tensor
-
-func (e experts) WriteTo(w io.Writer) (int64, error) {
-	// TODO(mxyng): experts _should_ be numerically sorted by expert but this should check
-	for _, t := range e {
-		// the canonical merged experts tensor stacks all experts along a new, 0 axis,
-		// e.g. `tensor.Stack(0, e[0], e[1:]...)`, which requires allocating temporary buffers
-		// this accomplishes the same thing by writing each expert tensor in sequence
-		if _, err := t.WriteTo(w); err != nil {
-			return 0, err
-		}
-	}
-
-	return 0, nil
-}

convert/convert_test.go

@@ -1,126 +0,0 @@
-package convert
-
-import (
-	"crypto/sha256"
-	"encoding/json"
-	"flag"
-	"fmt"
-	"io"
-	"io/fs"
-	"log/slog"
-	"math"
-	"os"
-	"path/filepath"
-	"slices"
-	"testing"
-
-	"github.com/ollama/ollama/llm"
-	"golang.org/x/exp/maps"
-)
-
-func convertFull(t *testing.T, fsys fs.FS) (*os.File, llm.KV, llm.Tensors) {
-	t.Helper()
-
-	f, err := os.CreateTemp(t.TempDir(), "f16")
-	if err != nil {
-		t.Fatal(err)
-	}
-	defer f.Close()
-
-	if err := Convert(fsys, f); err != nil {
-		t.Fatal(err)
-	}
-
-	r, err := os.Open(f.Name())
-	if err != nil {
-		t.Fatal(err)
-	}
-	t.Cleanup(func() { r.Close() })
-
-	m, _, err := llm.DecodeGGML(r, math.MaxInt)
-	if err != nil {
-		t.Fatal(err)
-	}
-
-	if _, err := r.Seek(0, io.SeekStart); err != nil {
-		t.Fatal(err)
-	}
-
-	return r, m.KV(), m.Tensors()
-}
-
-func TestMain(m *testing.M) {
-	var level slog.Level
-	flag.TextVar(&level, "level", slog.LevelInfo, "log level")
-	flag.Parse()
-	slog.SetLogLoggerLevel(level)
-	os.Exit(m.Run())
-}
-
-func TestConvertFull(t *testing.T) {
-	cases := []string{
-		"Meta-Llama-3-8B-Instruct",
-		"Mistral-7B-Instruct-v0.2",
-		"Mixtral-8x7B-Instruct-v0.1",
-		"gemma-2b-it",
-	}
-
-	for i := range cases {
-		tt := cases[i]
-		t.Run(tt, func(t *testing.T) {
-			t.Parallel()
-
-			p := filepath.Join("testdata", tt)
-			if testing.Short() {
-				t.Skip("skipping in short mode")
-			} else if _, err := os.Stat(p); err != nil {
-				t.Skipf("%s not found", p)
-			}
-
-			f, kv, tensors := convertFull(t, os.DirFS(p))
-			actual := make(map[string]string)
-			for k, v := range kv {
-				if s, ok := v.(json.Marshaler); !ok {
-					actual[k] = fmt.Sprintf("%v", v)
-				} else {
-					bts, err := json.Marshal(s)
-					if err != nil {
-						t.Fatal(err)
-					}
-
-					actual[k] = fmt.Sprintf("%x", sha256.Sum256(bts))
-				}
-			}
-
-			for _, tensor := range tensors.Items {
-				sha256sum := sha256.New()
-				sr := io.NewSectionReader(f, int64(tensors.Offset+tensor.Offset), int64(tensor.Size()))
-				if _, err := io.Copy(sha256sum, sr); err != nil {
-					t.Fatal(err)
-				}
-
-				actual[tensor.Name] = fmt.Sprintf("%x", sha256sum.Sum(nil))
-			}
-
-			expectFile, err := os.Open(filepath.Join("testdata", fmt.Sprintf("%s.json", tt)))
-			if err != nil {
-				t.Fatal(err)
-			}
-
-			var expect map[string]string
-			if err := json.NewDecoder(expectFile).Decode(&expect); err != nil {
-				t.Fatal(err)
-			}
-
-			keys := maps.Keys(expect)
-			slices.Sort(keys)
-			for _, k := range keys {
-				if v, ok := actual[k]; !ok {
-					t.Errorf("missing %s", k)
-				} else if v != expect[k] {
-					t.Errorf("unexpected %s: want %s, got %s", k, expect[k], v)
-				}
-			}
-		})
-	}
-}

convert/fs.go

@@ -1,58 +0,0 @@
-package convert
-
-import (
-	"archive/zip"
-	"errors"
-	"io"
-	"io/fs"
-	"os"
-	"path/filepath"
-)
-
-type ZipReader struct {
-	r     *zip.Reader
-	p     string
-
-	// limit is the maximum size of a file that can be read directly
-	// from the zip archive. Files larger than this size will be extracted
-	limit int64
-}
-
-func NewZipReader(r *zip.Reader, p string, limit int64) fs.FS {
-	return &ZipReader{r, p, limit}
-}
-
-func (z *ZipReader) Open(name string) (fs.File, error) {
-	r, err := z.r.Open(name)
-	if err != nil {
-		return nil, err
-	}
-	defer r.Close()
-
-	if fi, err := r.Stat(); err != nil {
-		return nil, err
-	} else if fi.Size() < z.limit {
-		return r, nil
-	}
-
-	if !filepath.IsLocal(name) {
-		return nil, zip.ErrInsecurePath
-	}
-
-	n := filepath.Join(z.p, name)
-	if _, err := os.Stat(n); errors.Is(err, os.ErrNotExist) {
-		w, err := os.Create(n)
-		if err != nil {
-			return nil, err
-		}
-		defer w.Close()
-
-		if _, err := io.Copy(w, r); err != nil {
-			return nil, err
-		}
-	} else if err != nil {
-		return nil, err
-	}
-
-	return os.Open(n)
-}