2026-01-19 17:35:25 +00:00
5 changed files with 684 additions and 1 deletions
--- a/.gitea/workflows/ci.yaml
+++ b/.gitea/workflows/ci.yaml
@@ -9,7 +9,6 @@ on:
  push:
    branches:
      - main
-      - gpu_support

 env:
  PADDLE_VERSION: "3.0.0"
@@ -24,7 +23,9 @@ jobs:
      image_cpu: seryus.ddns.net/unir/paddle-ocr-cpu
      image_gpu: seryus.ddns.net/unir/paddle-ocr-gpu
      image_easyocr: seryus.ddns.net/unir/easyocr-cpu
+      image_easyocr_gpu: seryus.ddns.net/unir/easyocr-gpu
      image_doctr: seryus.ddns.net/unir/doctr-cpu
+      image_doctr_gpu: seryus.ddns.net/unir/doctr-gpu
    steps:
      - name: Output version info
        run: |
@@ -315,3 +316,139 @@ jobs:
          docker buildx imagetools create -t ${{ needs.essential.outputs.image_doctr }}:${{ needs.essential.outputs.Version }} \
            ${{ needs.essential.outputs.image_doctr }}:${{ needs.essential.outputs.Version }}-amd64 \
            ${{ needs.essential.outputs.image_doctr }}:${{ needs.essential.outputs.Version }}-arm64
+
+  # EasyOCR GPU image: Matrix build for amd64 and arm64
+  # PyTorch cu128 has wheels for both architectures
+  build_easyocr_gpu:
+    runs-on: ubuntu-latest
+    needs: essential
+    strategy:
+      matrix:
+        platform:
+          - linux/amd64
+          - linux/arm64
+    steps:
+      - name: Checkout
+        uses: actions/checkout@v4
+
+      - name: Set up QEMU
+        uses: docker/setup-qemu-action@v3
+
+      - name: Set up Docker Buildx
+        uses: docker/setup-buildx-action@v3
+
+      - name: Login to Gitea Registry
+        uses: docker/login-action@v3
+        with:
+          registry: ${{ needs.essential.outputs.repo }}
+          username: username
+          password: ${{ secrets.CI_READWRITE }}
+
+      - name: Get arch suffix
+        id: arch
+        run: |
+          if [ "${{ matrix.platform }}" = "linux/amd64" ]; then
+            echo "suffix=amd64" >> $GITHUB_OUTPUT
+          else
+            echo "suffix=arm64" >> $GITHUB_OUTPUT
+          fi
+
+      - name: Build and push EasyOCR GPU image (${{ matrix.platform }})
+        uses: docker/build-push-action@v5
+        with:
+          context: src/easyocr_service
+          file: src/easyocr_service/Dockerfile.gpu
+          platforms: ${{ matrix.platform }}
+          push: true
+          tags: |
+            ${{ needs.essential.outputs.image_easyocr_gpu }}:${{ needs.essential.outputs.Version }}-${{ steps.arch.outputs.suffix }}
+            ${{ needs.essential.outputs.image_easyocr_gpu }}:${{ steps.arch.outputs.suffix }}
+
+  # DocTR GPU image: Matrix build for amd64 and arm64
+  # PyTorch cu128 has wheels for both architectures
+  build_doctr_gpu:
+    runs-on: ubuntu-latest
+    needs: essential
+    strategy:
+      matrix:
+        platform:
+          - linux/amd64
+          - linux/arm64
+    steps:
+      - name: Checkout
+        uses: actions/checkout@v4
+
+      - name: Set up QEMU
+        uses: docker/setup-qemu-action@v3
+
+      - name: Set up Docker Buildx
+        uses: docker/setup-buildx-action@v3
+
+      - name: Login to Gitea Registry
+        uses: docker/login-action@v3
+        with:
+          registry: ${{ needs.essential.outputs.repo }}
+          username: username
+          password: ${{ secrets.CI_READWRITE }}
+
+      - name: Get arch suffix
+        id: arch
+        run: |
+          if [ "${{ matrix.platform }}" = "linux/amd64" ]; then
+            echo "suffix=amd64" >> $GITHUB_OUTPUT
+          else
+            echo "suffix=arm64" >> $GITHUB_OUTPUT
+          fi
+
+      - name: Build and push DocTR GPU image (${{ matrix.platform }})
+        uses: docker/build-push-action@v5
+        with:
+          context: src/doctr_service
+          file: src/doctr_service/Dockerfile.gpu
+          platforms: ${{ matrix.platform }}
+          push: true
+          tags: |
+            ${{ needs.essential.outputs.image_doctr_gpu }}:${{ needs.essential.outputs.Version }}-${{ steps.arch.outputs.suffix }}
+            ${{ needs.essential.outputs.image_doctr_gpu }}:${{ steps.arch.outputs.suffix }}
+
+  # Create multi-arch manifest for EasyOCR GPU image
+  manifest_easyocr_gpu:
+    runs-on: ubuntu-latest
+    needs: [essential, build_easyocr_gpu]
+    steps:
+      - name: Login to Gitea Registry
+        uses: docker/login-action@v3
+        with:
+          registry: ${{ needs.essential.outputs.repo }}
+          username: username
+          password: ${{ secrets.CI_READWRITE }}
+
+      - name: Create multi-arch manifest (EasyOCR GPU)
+        run: |
+          docker buildx imagetools create -t ${{ needs.essential.outputs.image_easyocr_gpu }}:latest \
+            ${{ needs.essential.outputs.image_easyocr_gpu }}:amd64 \
+            ${{ needs.essential.outputs.image_easyocr_gpu }}:arm64
+          docker buildx imagetools create -t ${{ needs.essential.outputs.image_easyocr_gpu }}:${{ needs.essential.outputs.Version }} \
+            ${{ needs.essential.outputs.image_easyocr_gpu }}:${{ needs.essential.outputs.Version }}-amd64 \
+            ${{ needs.essential.outputs.image_easyocr_gpu }}:${{ needs.essential.outputs.Version }}-arm64
+
+  # Create multi-arch manifest for DocTR GPU image
+  manifest_doctr_gpu:
+    runs-on: ubuntu-latest
+    needs: [essential, build_doctr_gpu]
+    steps:
+      - name: Login to Gitea Registry
+        uses: docker/login-action@v3
+        with:
+          registry: ${{ needs.essential.outputs.repo }}
+          username: username
+          password: ${{ secrets.CI_READWRITE }}
+
+      - name: Create multi-arch manifest (DocTR GPU)
+        run: |
+          docker buildx imagetools create -t ${{ needs.essential.outputs.image_doctr_gpu }}:latest \
+            ${{ needs.essential.outputs.image_doctr_gpu }}:amd64 \
+            ${{ needs.essential.outputs.image_doctr_gpu }}:arm64
+          docker buildx imagetools create -t ${{ needs.essential.outputs.image_doctr_gpu }}:${{ needs.essential.outputs.Version }} \
+            ${{ needs.essential.outputs.image_doctr_gpu }}:${{ needs.essential.outputs.Version }}-amd64 \
+            ${{ needs.essential.outputs.image_doctr_gpu }}:${{ needs.essential.outputs.Version }}-arm64
--- a/src/doctr_service/Dockerfile.gpu
+++ b/src/doctr_service/Dockerfile.gpu
@@ -0,0 +1,68 @@
+# Dockerfile.gpu - DocTR GPU Dockerfile for amd64/arm64
+#
+# Build:
+#   docker build -t doctr-gpu:latest -f Dockerfile.gpu .
+#
+# Run:
+#   docker run --gpus all -p 8003:8000 -v ./dataset:/app/dataset doctr-gpu:latest
+
+# CUDA 13.0 for Blackwell (sm_121) and GH200/GB200 support
+FROM nvidia/cuda:13.0.2-cudnn-runtime-ubuntu24.04
+
+LABEL maintainer="Sergio Jimenez"
+LABEL description="DocTR Tuning REST API - GPU/CUDA version"
+
+WORKDIR /app
+
+# Set environment variables
+ENV DEBIAN_FRONTEND=noninteractive
+ENV PYTHONUNBUFFERED=1
+ENV CUDA_VISIBLE_DEVICES=0
+ENV DOCTR_DET_ARCH=db_resnet50
+ENV DOCTR_RECO_ARCH=crnn_vgg16_bn
+
+# Install Python 3.12 and system dependencies
+RUN apt-get update && apt-get install -y --no-install-recommends \
+    python3.12 \
+    python3.12-venv \
+    python3-pip \
+    libgl1 \
+    libglib2.0-0 \
+    libsm6 \
+    libxext6 \
+    libxrender1 \
+    libgomp1 \
+    && rm -rf /var/lib/apt/lists/* \
+    && ln -sf /usr/bin/python3.12 /usr/bin/python
+
+# Install PyTorch with CUDA support
+# cu128 index has both amd64 and arm64 wheels
+RUN python -m pip install --no-cache-dir \
+    torch torchvision --index-url https://download.pytorch.org/whl/cu128
+
+# Install DocTR and other dependencies
+RUN python -m pip install --no-cache-dir \
+    "python-doctr[torch]>=0.8.0" \
+    fastapi>=0.104.0 \
+    "uvicorn[standard]" \
+    pydantic>=2.0.0 \
+    jiwer>=3.0.0 \
+    numpy>=1.24.0 \
+    pillow>=10.0.0
+
+# Copy application code
+COPY doctr_tuning_rest.py .
+COPY dataset_manager.py .
+
+# Volume for dataset and model cache
+VOLUME ["/app/dataset", "/root/.cache/doctr"]
+
+# Expose API port
+EXPOSE 8000
+
+# Health check (longer start period for model download)
+HEALTHCHECK --interval=30s --timeout=10s --start-period=180s --retries=3 \
+    CMD python -c "import urllib.request; urllib.request.urlopen('http://localhost:8000/health')" || exit 1
+
+# Run the API server
+CMD ["uvicorn", "doctr_tuning_rest:app", "--host", "0.0.0.0", "--port", "8000"]
--- a/src/doctr_service/README.md
+++ b/src/doctr_service/README.md
@@ -0,0 +1,212 @@
+# DocTR Tuning REST API
+
+REST API service for DocTR (Document Text Recognition) hyperparameter evaluation. Keeps the model loaded in memory for fast repeated evaluations during hyperparameter search.
+
+## Quick Start
+
+### CPU Version
+
+```bash
+cd src/doctr_service
+
+# Build
+docker build -t doctr-api:cpu .
+
+# Run
+docker run -d -p 8003:8000 \
+  -v $(pwd)/../dataset:/app/dataset:ro \
+  -v doctr-cache:/root/.cache/doctr \
+  doctr-api:cpu
+
+# Test
+curl http://localhost:8003/health
+```
+
+### GPU Version
+
+```bash
+# Build GPU image
+docker build -f Dockerfile.gpu -t doctr-api:gpu .
+
+# Run with GPU
+docker run -d -p 8003:8000 --gpus all \
+  -v $(pwd)/../dataset:/app/dataset:ro \
+  -v doctr-cache:/root/.cache/doctr \
+  doctr-api:gpu
+```
+
+## Files
+
+| File | Description |
+|------|-------------|
+| `doctr_tuning_rest.py` | FastAPI REST service with 9 tunable hyperparameters |
+| `dataset_manager.py` | Dataset loader (shared with other services) |
+| `Dockerfile` | CPU-only image (amd64 + arm64) |
+| `Dockerfile.gpu` | GPU/CUDA image (amd64 + arm64) |
+| `requirements.txt` | Python dependencies |
+
+## API Endpoints
+
+### `GET /health`
+
+Check if service is ready.
+
+```json
+{
+  "status": "ok",
+  "model_loaded": true,
+  "dataset_loaded": true,
+  "dataset_size": 24,
+  "det_arch": "db_resnet50",
+  "reco_arch": "crnn_vgg16_bn",
+  "cuda_available": true,
+  "device": "cuda",
+  "gpu_name": "NVIDIA GB10"
+}
+```
+
+### `POST /evaluate`
+
+Run OCR evaluation with given hyperparameters.
+
+**Request (9 tunable parameters):**
+```json
+{
+  "pdf_folder": "/app/dataset",
+  "assume_straight_pages": true,
+  "straighten_pages": false,
+  "preserve_aspect_ratio": true,
+  "symmetric_pad": true,
+  "disable_page_orientation": false,
+  "disable_crop_orientation": false,
+  "resolve_lines": true,
+  "resolve_blocks": false,
+  "paragraph_break": 0.035,
+  "start_page": 5,
+  "end_page": 10
+}
+```
+
+**Response:**
+```json
+{
+  "CER": 0.0189,
+  "WER": 0.1023,
+  "TIME": 52.3,
+  "PAGES": 5,
+  "TIME_PER_PAGE": 10.46,
+  "model_reinitialized": false
+}
+```
+
+**Note:** `model_reinitialized` indicates if the model was reloaded due to changed processing flags (adds ~2-5s overhead).
+
+## Hyperparameters
+
+### Processing Flags (Require Model Reinitialization)
+
+| Parameter | Default | Description |
+|-----------|---------|-------------|
+| `assume_straight_pages` | true | Skip rotation handling for straight documents |
+| `straighten_pages` | false | Pre-straighten pages before detection |
+| `preserve_aspect_ratio` | true | Maintain document proportions during resize |
+| `symmetric_pad` | true | Use symmetric padding when preserving aspect ratio |
+
+**Note:** Changing these flags requires model reinitialization (~2-5s).
+
+### Orientation Flags
+
+| Parameter | Default | Description |
+|-----------|---------|-------------|
+| `disable_page_orientation` | false | Skip page orientation classification |
+| `disable_crop_orientation` | false | Skip crop orientation detection |
+
+### Output Grouping
+
+| Parameter | Default | Range | Description |
+|-----------|---------|-------|-------------|
+| `resolve_lines` | true | bool | Group words into lines |
+| `resolve_blocks` | false | bool | Group lines into blocks |
+| `paragraph_break` | 0.035 | 0.0-1.0 | Minimum space ratio separating paragraphs |
+
+## Model Architecture
+
+DocTR uses a two-stage pipeline:
+
+1. **Detection** (`det_arch`): Localizes text regions
+   - Default: `db_resnet50` (DBNet with ResNet-50 backbone)
+   - Alternatives: `linknet_resnet18`, `db_mobilenet_v3_large`
+
+2. **Recognition** (`reco_arch`): Recognizes characters
+   - Default: `crnn_vgg16_bn` (CRNN with VGG-16 backbone)
+   - Alternatives: `sar_resnet31`, `master`, `vitstr_small`
+
+Architecture is set via environment variables (fixed at startup).
+
+## GPU Support
+
+### Platform Support
+
+| Platform | CPU | GPU |
+|----------|-----|-----|
+| Linux x86_64 (amd64) | ✅ | ✅ PyTorch CUDA |
+| Linux ARM64 (GH200/GB200/DGX Spark) | ✅ | ✅ PyTorch CUDA (cu128 index) |
+| macOS ARM64 (M1/M2) | ✅ | ❌ |
+
+### PyTorch CUDA on ARM64
+
+Unlike PaddlePaddle, PyTorch provides **official ARM64 CUDA wheels** on the cu128 index:
+
+```bash
+pip install torch torchvision --index-url https://download.pytorch.org/whl/cu128
+```
+
+This works on both amd64 and arm64 platforms with CUDA support.
+
+### GPU Detection
+
+DocTR automatically uses GPU when available:
+
+```python
+import torch
+print(torch.cuda.is_available())  # True if GPU available
+
+# DocTR model moves to GPU
+model = ocr_predictor(pretrained=True)
+if torch.cuda.is_available():
+    model = model.cuda()
+```
+
+The `/health` endpoint shows GPU status:
+```json
+{
+  "cuda_available": true,
+  "device": "cuda",
+  "gpu_name": "NVIDIA GB10",
+  "gpu_memory_total": "128.00 GB"
+}
+```
+
+## Environment Variables
+
+| Variable | Default | Description |
+|----------|---------|-------------|
+| `DOCTR_DET_ARCH` | `db_resnet50` | Detection architecture |
+| `DOCTR_RECO_ARCH` | `crnn_vgg16_bn` | Recognition architecture |
+| `CUDA_VISIBLE_DEVICES` | `0` | GPU device selection |
+
+## CI/CD
+
+Built images available from registry:
+
+| Image | Architecture |
+|-------|--------------|
+| `seryus.ddns.net/unir/doctr-cpu:latest` | amd64, arm64 |
+| `seryus.ddns.net/unir/doctr-gpu:latest` | amd64, arm64 |
+
+## Sources
+
+- [DocTR Documentation](https://mindee.github.io/doctr/)
+- [DocTR GitHub](https://github.com/mindee/doctr)
+- [DocTR Model Usage](https://mindee.github.io/doctr/latest/using_doctr/using_models.html)
+- [PyTorch ARM64 CUDA Wheels](https://github.com/pytorch/pytorch/issues/160162)
--- a/src/easyocr_service/Dockerfile.gpu
+++ b/src/easyocr_service/Dockerfile.gpu
@@ -0,0 +1,67 @@
+# Dockerfile.gpu - EasyOCR GPU Dockerfile for amd64/arm64
+#
+# Build:
+#   docker build -t easyocr-gpu:latest -f Dockerfile.gpu .
+#
+# Run:
+#   docker run --gpus all -p 8002:8000 -v ./dataset:/app/dataset easyocr-gpu:latest
+
+# CUDA 13.0 for Blackwell (sm_121) and GH200/GB200 support
+FROM nvidia/cuda:13.0.2-cudnn-runtime-ubuntu24.04
+
+LABEL maintainer="Sergio Jimenez"
+LABEL description="EasyOCR Tuning REST API - GPU/CUDA version"
+
+WORKDIR /app
+
+# Set environment variables
+ENV DEBIAN_FRONTEND=noninteractive
+ENV PYTHONUNBUFFERED=1
+ENV CUDA_VISIBLE_DEVICES=0
+ENV EASYOCR_LANGUAGES=es,en
+
+# Install Python 3.12 and system dependencies
+RUN apt-get update && apt-get install -y --no-install-recommends \
+    python3.12 \
+    python3.12-venv \
+    python3-pip \
+    libgl1 \
+    libglib2.0-0 \
+    libsm6 \
+    libxext6 \
+    libxrender1 \
+    libgomp1 \
+    && rm -rf /var/lib/apt/lists/* \
+    && ln -sf /usr/bin/python3.12 /usr/bin/python
+
+# Install PyTorch with CUDA support
+# cu128 index has both amd64 and arm64 wheels
+RUN python -m pip install --no-cache-dir \
+    torch torchvision --index-url https://download.pytorch.org/whl/cu128
+
+# Install EasyOCR and other dependencies
+RUN python -m pip install --no-cache-dir \
+    easyocr>=1.7.0 \
+    fastapi>=0.104.0 \
+    "uvicorn[standard]" \
+    pydantic>=2.0.0 \
+    jiwer>=3.0.0 \
+    numpy>=1.24.0 \
+    pillow>=10.0.0
+
+# Copy application code
+COPY easyocr_tuning_rest.py .
+COPY dataset_manager.py .
+
+# Volume for dataset and model cache
+VOLUME ["/app/dataset", "/root/.EasyOCR"]
+
+# Expose API port
+EXPOSE 8000
+
+# Health check
+HEALTHCHECK --interval=30s --timeout=10s --start-period=120s --retries=3 \
+    CMD python -c "import urllib.request; urllib.request.urlopen('http://localhost:8000/health')" || exit 1
+
+# Run the API server
+CMD ["uvicorn", "easyocr_tuning_rest:app", "--host", "0.0.0.0", "--port", "8000"]
--- a/src/easyocr_service/README.md
+++ b/src/easyocr_service/README.md
@@ -0,0 +1,199 @@
+# EasyOCR Tuning REST API
+
+REST API service for EasyOCR hyperparameter evaluation. Keeps the model loaded in memory for fast repeated evaluations during hyperparameter search.
+
+## Quick Start
+
+### CPU Version
+
+```bash
+cd src/easyocr_service
+
+# Build
+docker build -t easyocr-api:cpu .
+
+# Run
+docker run -d -p 8002:8000 \
+  -v $(pwd)/../dataset:/app/dataset:ro \
+  -v easyocr-cache:/root/.EasyOCR \
+  easyocr-api:cpu
+
+# Test
+curl http://localhost:8002/health
+```
+
+### GPU Version
+
+```bash
+# Build GPU image
+docker build -f Dockerfile.gpu -t easyocr-api:gpu .
+
+# Run with GPU
+docker run -d -p 8002:8000 --gpus all \
+  -v $(pwd)/../dataset:/app/dataset:ro \
+  -v easyocr-cache:/root/.EasyOCR \
+  easyocr-api:gpu
+```
+
+## Files
+
+| File | Description |
+|------|-------------|
+| `easyocr_tuning_rest.py` | FastAPI REST service with 14 tunable hyperparameters |
+| `dataset_manager.py` | Dataset loader (shared with other services) |
+| `Dockerfile` | CPU-only image (amd64 + arm64) |
+| `Dockerfile.gpu` | GPU/CUDA image (amd64 + arm64) |
+| `requirements.txt` | Python dependencies |
+
+## API Endpoints
+
+### `GET /health`
+
+Check if service is ready.
+
+```json
+{
+  "status": "ok",
+  "model_loaded": true,
+  "dataset_loaded": true,
+  "dataset_size": 24,
+  "languages": ["es", "en"],
+  "cuda_available": true,
+  "device": "cuda",
+  "gpu_name": "NVIDIA GB10"
+}
+```
+
+### `POST /evaluate`
+
+Run OCR evaluation with given hyperparameters.
+
+**Request (14 tunable parameters):**
+```json
+{
+  "pdf_folder": "/app/dataset",
+  "text_threshold": 0.7,
+  "low_text": 0.4,
+  "link_threshold": 0.4,
+  "slope_ths": 0.1,
+  "ycenter_ths": 0.5,
+  "height_ths": 0.5,
+  "width_ths": 0.5,
+  "add_margin": 0.1,
+  "contrast_ths": 0.1,
+  "adjust_contrast": 0.5,
+  "decoder": "greedy",
+  "beamWidth": 5,
+  "min_size": 10,
+  "rotation_info": null,
+  "start_page": 5,
+  "end_page": 10
+}
+```
+
+**Response:**
+```json
+{"CER": 0.0234, "WER": 0.1156, "TIME": 45.2, "PAGES": 5, "TIME_PER_PAGE": 9.04}
+```
+
+## Hyperparameters
+
+### Detection (CRAFT Algorithm)
+
+| Parameter | Default | Range | Description |
+|-----------|---------|-------|-------------|
+| `text_threshold` | 0.7 | 0.0-1.0 | Text confidence threshold |
+| `low_text` | 0.4 | 0.0-1.0 | Text lower-bound score |
+| `link_threshold` | 0.4 | 0.0-1.0 | Link confidence threshold |
+
+### Bounding Box Merging
+
+| Parameter | Default | Range | Description |
+|-----------|---------|-------|-------------|
+| `slope_ths` | 0.1 | 0.0-1.0 | Max slope for merging |
+| `ycenter_ths` | 0.5 | 0.0-2.0 | Max vertical shift |
+| `height_ths` | 0.5 | 0.0-2.0 | Max height variance |
+| `width_ths` | 0.5 | 0.0-2.0 | Max horizontal distance |
+| `add_margin` | 0.1 | 0.0-1.0 | Bounding box extension |
+
+### Contrast
+
+| Parameter | Default | Range | Description |
+|-----------|---------|-------|-------------|
+| `contrast_ths` | 0.1 | 0.0-1.0 | Contrast threshold for dual-pass |
+| `adjust_contrast` | 0.5 | 0.0-1.0 | Target contrast level |
+
+### Decoder
+
+| Parameter | Default | Options | Description |
+|-----------|---------|---------|-------------|
+| `decoder` | "greedy" | greedy, beamsearch, wordbeamsearch | Decoding method |
+| `beamWidth` | 5 | 1-20 | Beam width (for beam search) |
+
+### Other
+
+| Parameter | Default | Description |
+|-----------|---------|-------------|
+| `min_size` | 10 | Minimum text box pixels |
+| `rotation_info` | null | Rotation angles to try: [90, 180, 270] |
+
+## GPU Support
+
+### Platform Support
+
+| Platform | CPU | GPU |
+|----------|-----|-----|
+| Linux x86_64 (amd64) | ✅ | ✅ PyTorch CUDA |
+| Linux ARM64 (GH200/GB200/DGX Spark) | ✅ | ✅ PyTorch CUDA (cu128 index) |
+| macOS ARM64 (M1/M2) | ✅ | ❌ |
+
+### PyTorch CUDA on ARM64
+
+Unlike PaddlePaddle, PyTorch provides **official ARM64 CUDA wheels** on the cu128 index:
+
+```bash
+pip install torch torchvision --index-url https://download.pytorch.org/whl/cu128
+```
+
+This works on both amd64 and arm64 platforms with CUDA support.
+
+### GPU Detection
+
+EasyOCR automatically uses GPU when PyTorch CUDA is available:
+
+```python
+import torch
+print(torch.cuda.is_available())  # True if GPU available
+```
+
+The `/health` endpoint shows GPU status:
+```json
+{
+  "cuda_available": true,
+  "device": "cuda",
+  "gpu_name": "NVIDIA GB10",
+  "gpu_memory_total": "128.00 GB"
+}
+```
+
+## Environment Variables
+
+| Variable | Default | Description |
+|----------|---------|-------------|
+| `EASYOCR_LANGUAGES` | `es,en` | Comma-separated language codes |
+| `CUDA_VISIBLE_DEVICES` | `0` | GPU device selection |
+
+## CI/CD
+
+Built images available from registry:
+
+| Image | Architecture |
+|-------|--------------|
+| `seryus.ddns.net/unir/easyocr-cpu:latest` | amd64, arm64 |
+| `seryus.ddns.net/unir/easyocr-gpu:latest` | amd64, arm64 |
+
+## Sources
+
+- [EasyOCR Documentation](https://www.jaided.ai/easyocr/documentation/)
+- [EasyOCR GitHub](https://github.com/JaidedAI/EasyOCR)
+- [PyTorch ARM64 CUDA Wheels](https://github.com/pytorch/pytorch/issues/160162)