feat: add GPU/CUDA acceleration support (:cuda Docker tag)

Add a :cuda Docker image tag that auto-detects NVIDIA GPU at runtime
and falls back gracefully to CPU. Same pattern as Immich.

- New gpu.py shared utility for cached CUDA detection
- Background removal (rembg): pass CUDAExecutionProvider to ONNX Runtime
- Upscaling (Real-ESRGAN): use CUDA device + FP16 when GPU available
- OCR (PaddleOCR): enable use_gpu when CUDA detected
- Dispatcher reports GPU status at startup via readiness signal
- Admin health endpoint exposes GPU availability
- Dockerfile uses ARG GPU=false with conditional NVIDIA CUDA base image
- docker-compose.gpu.yml override for GPU users
- CI/CD workflows build and publish :cuda tag (amd64 only)

Three tags: :latest (CPU), :lite (no AI), :cuda (GPU with CPU fallback)

.github/workflows/ci.yml

@@ -80,11 +80,20 @@ jobs:
       - run: pnpm build
 
   docker:
-    name: Docker Build Test (${{ matrix.variant }})
+    name: Docker Build Test (${{ matrix.tag }})
     runs-on: ubuntu-latest
     strategy:
       matrix:
-        variant: [full, lite]
+        include:
+          - tag: full
+            variant: full
+            gpu: "false"
+          - tag: lite
+            variant: lite
+            gpu: "false"
+          - tag: cuda
+            variant: full
+            gpu: "true"
     steps:
       - uses: actions/checkout@v4
 
@@ -95,7 +104,9 @@ jobs:
           context: .
           file: docker/Dockerfile
           push: false
-          build-args: VARIANT=${{ matrix.variant }}
-          tags: stirling-image:ci-${{ matrix.variant }}
-          cache-from: type=gha,scope=${{ matrix.variant }}
-          cache-to: type=gha,mode=max,scope=${{ matrix.variant }}
+          build-args: |
+            VARIANT=${{ matrix.variant }}
+            GPU=${{ matrix.gpu }}
+          tags: stirling-image:ci-${{ matrix.tag }}
+          cache-from: type=gha,scope=${{ matrix.tag }}
+          cache-to: type=gha,mode=max,scope=${{ matrix.tag }}

.github/workflows/release.yml

@@ -49,18 +49,28 @@ jobs:
           fi
 
   docker:
-    name: Docker (${{ matrix.variant }})
+    name: Docker (${{ matrix.tag }})
     needs: release
     if: needs.release.outputs.new_version != ''
     runs-on: ubuntu-latest
     strategy:
       matrix:
-        variant: [full, lite]
         include:
-          - variant: full
+          - tag: full
+            variant: full
+            gpu: "false"
             suffix: ""
-          - variant: lite
+            platforms: "linux/amd64,linux/arm64"
+          - tag: lite
+            variant: lite
+            gpu: "false"
             suffix: "-lite"
+            platforms: "linux/amd64,linux/arm64"
+          - tag: cuda
+            variant: full
+            gpu: "true"
+            suffix: "-cuda"
+            platforms: "linux/amd64"
     steps:
       - name: Checkout release tag
         uses: actions/checkout@v4
@@ -97,7 +107,7 @@ jobs:
             type=semver,pattern={{version}}${{ matrix.suffix }},value=v${{ needs.release.outputs.new_version }}
             type=semver,pattern={{major}}.{{minor}}${{ matrix.suffix }},value=v${{ needs.release.outputs.new_version }}
             type=semver,pattern={{major}}${{ matrix.suffix }},value=v${{ needs.release.outputs.new_version }}
-            type=raw,value=${{ matrix.variant == 'full' && 'latest' || 'lite' }}
+            type=raw,value=${{ matrix.tag == 'full' && 'latest' || matrix.tag }}
 
       - name: Build and push
         uses: docker/build-push-action@v6
@@ -105,9 +115,11 @@ jobs:
           context: .
           file: docker/Dockerfile
           push: true
-          build-args: VARIANT=${{ matrix.variant }}
-          platforms: linux/amd64,linux/arm64
+          build-args: |
+            VARIANT=${{ matrix.variant }}
+            GPU=${{ matrix.gpu }}
+          platforms: ${{ matrix.platforms }}
           tags: ${{ steps.meta.outputs.tags }}
           labels: ${{ steps.meta.outputs.labels }}
-          cache-from: type=gha,scope=${{ matrix.variant }}
-          cache-to: type=gha,mode=max,scope=${{ matrix.variant }}
+          cache-from: type=gha,scope=${{ matrix.tag }}
+          cache-to: type=gha,mode=max,scope=${{ matrix.tag }}