fix(docker): standardize Dockerfile and entrypoint filenames; add GPU-specific Dockerfiles and update build and test references

Dockerfile_minicpm45v_cpu

@@ -14,7 +14,7 @@ ENV OLLAMA_ORIGINS="*"
 ENV CUDA_VISIBLE_DEVICES=""
 
 # Copy and setup entrypoint
-COPY image_support_files/docker-entrypoint.sh /usr/local/bin/docker-entrypoint.sh
+COPY image_support_files/minicpm45v_entrypoint.sh /usr/local/bin/docker-entrypoint.sh
 RUN chmod +x /usr/local/bin/docker-entrypoint.sh
 
 # Expose Ollama API port

Dockerfile_minicpm45v_gpu

@@ -12,7 +12,7 @@ ENV OLLAMA_HOST="0.0.0.0"
 ENV OLLAMA_ORIGINS="*"
 
 # Copy and setup entrypoint
-COPY image_support_files/docker-entrypoint.sh /usr/local/bin/docker-entrypoint.sh
+COPY image_support_files/minicpm45v_entrypoint.sh /usr/local/bin/docker-entrypoint.sh
 RUN chmod +x /usr/local/bin/docker-entrypoint.sh
 
 # Expose Ollama API port

Dockerfile_paddleocr_vl

@@ -1,70 +0,0 @@
-# PaddleOCR-VL GPU Variant
-# Vision-Language Model for document parsing using vLLM
-FROM nvidia/cuda:12.4.0-devel-ubuntu22.04
-
-LABEL maintainer="Task Venture Capital GmbH <hello@task.vc>"
-LABEL description="PaddleOCR-VL 0.9B - Vision-Language Model for document parsing"
-LABEL org.opencontainers.image.source="https://code.foss.global/host.today/ht-docker-ai"
-
-# Environment configuration
-ENV DEBIAN_FRONTEND=noninteractive
-ENV PYTHONUNBUFFERED=1
-ENV HF_HOME=/root/.cache/huggingface
-ENV VLLM_WORKER_MULTIPROC_METHOD=spawn
-
-# Set working directory
-WORKDIR /app
-
-# Install system dependencies
-RUN apt-get update && apt-get install -y --no-install-recommends \
-    python3.11 \
-    python3.11-venv \
-    python3.11-dev \
-    python3-pip \
-    git \
-    curl \
-    build-essential \
-    && rm -rf /var/lib/apt/lists/* \
-    && update-alternatives --install /usr/bin/python python /usr/bin/python3.11 1 \
-    && update-alternatives --install /usr/bin/python3 python3 /usr/bin/python3.11 1
-
-# Create and activate virtual environment
-RUN python -m venv /opt/venv
-ENV PATH="/opt/venv/bin:$PATH"
-
-# Install PyTorch with CUDA support
-RUN pip install --no-cache-dir --upgrade pip && \
-    pip install --no-cache-dir \
-    torch==2.5.1 \
-    torchvision \
-    --index-url https://download.pytorch.org/whl/cu124
-
-# Install vLLM 0.11.1 (first stable release with PaddleOCR-VL support)
-RUN pip install --no-cache-dir \
-    vllm==0.11.1 \
-    --extra-index-url https://download.pytorch.org/whl/cu124
-
-# Install additional dependencies
-RUN pip install --no-cache-dir \
-    transformers \
-    accelerate \
-    safetensors \
-    pillow \
-    fastapi \
-    uvicorn[standard] \
-    python-multipart \
-    openai \
-    httpx
-
-# Copy entrypoint script
-COPY image_support_files/paddleocr-vl-entrypoint.sh /usr/local/bin/paddleocr-vl-entrypoint.sh
-RUN chmod +x /usr/local/bin/paddleocr-vl-entrypoint.sh
-
-# Expose vLLM API port
-EXPOSE 8000
-
-# Health check
-HEALTHCHECK --interval=30s --timeout=10s --start-period=300s --retries=3 \
-    CMD curl -f http://localhost:8000/health || exit 1
-
-ENTRYPOINT ["/usr/local/bin/paddleocr-vl-entrypoint.sh"]

Dockerfile_paddleocr_vl_cpu

@@ -44,7 +44,7 @@ RUN pip install --no-cache-dir --upgrade pip && \
 
 # Copy server files
 COPY image_support_files/paddleocr_vl_server.py /app/paddleocr_vl_server.py
-COPY image_support_files/paddleocr-vl-cpu-entrypoint.sh /usr/local/bin/paddleocr-vl-cpu-entrypoint.sh
+COPY image_support_files/paddleocr_vl_entrypoint.sh /usr/local/bin/paddleocr-vl-cpu-entrypoint.sh
 RUN chmod +x /usr/local/bin/paddleocr-vl-cpu-entrypoint.sh
 
 # Expose API port

Dockerfile_paddleocr_vl_gpu

@@ -58,7 +58,7 @@ RUN pip install --no-cache-dir \
 
 # Copy server files (same as CPU variant - it auto-detects CUDA)
 COPY image_support_files/paddleocr_vl_server.py /app/paddleocr_vl_server.py
-COPY image_support_files/paddleocr-vl-cpu-entrypoint.sh /usr/local/bin/paddleocr-vl-entrypoint.sh
+COPY image_support_files/paddleocr_vl_entrypoint.sh /usr/local/bin/paddleocr-vl-entrypoint.sh
 RUN chmod +x /usr/local/bin/paddleocr-vl-entrypoint.sh
 
 # Expose API port

build-images.sh

@@ -16,7 +16,7 @@ echo -e "${BLUE}Building ht-docker-ai images...${NC}"
 # Build GPU variant
 echo -e "${GREEN}Building MiniCPM-V 4.5 GPU variant...${NC}"
 docker build \
-    -f Dockerfile_minicpm45v \
+    -f Dockerfile_minicpm45v_gpu \
     -t ${REGISTRY}/${NAMESPACE}/${IMAGE_NAME}:minicpm45v \
     -t ${REGISTRY}/${NAMESPACE}/${IMAGE_NAME}:minicpm45v-gpu \
     -t ${REGISTRY}/${NAMESPACE}/${IMAGE_NAME}:latest \
@@ -29,10 +29,10 @@ docker build \
     -t ${REGISTRY}/${NAMESPACE}/${IMAGE_NAME}:minicpm45v-cpu \
     .
 
-# Build PaddleOCR-VL GPU variant (vLLM)
-echo -e "${GREEN}Building PaddleOCR-VL GPU variant (vLLM)...${NC}"
+# Build PaddleOCR-VL GPU variant
+echo -e "${GREEN}Building PaddleOCR-VL GPU variant...${NC}"
 docker build \
-    -f Dockerfile_paddleocr_vl \
+    -f Dockerfile_paddleocr_vl_gpu \
     -t ${REGISTRY}/${NAMESPACE}/${IMAGE_NAME}:paddleocr-vl \
     -t ${REGISTRY}/${NAMESPACE}/${IMAGE_NAME}:paddleocr-vl-gpu \
     .

changelog.md

@@ -1,5 +1,13 @@
 # Changelog
 
+## 2026-01-17 - 1.7.1 - fix(docker)
+standardize Dockerfile and entrypoint filenames; add GPU-specific Dockerfiles and update build and test references
+
+- Added Dockerfile_minicpm45v_gpu and image_support_files/minicpm45v_entrypoint.sh; removed the old Dockerfile_minicpm45v and docker-entrypoint.sh
+- Renamed and simplified PaddleOCR entrypoint to image_support_files/paddleocr_vl_entrypoint.sh and updated CPU/GPU Dockerfile references
+- Updated build-images.sh to use *_gpu Dockerfiles and clarified PaddleOCR GPU build log
+- Updated test/helpers/docker.ts to point to Dockerfile_minicpm45v_gpu so tests build the GPU variant
+
 ## 2026-01-17 - 1.7.0 - feat(tests)
 use Qwen2.5 (Ollama) for invoice extraction tests and add helpers for model management; normalize dates and coerce numeric fields
 

image_support_files/paddleocr-vl-entrypoint.sh

@@ -1,59 +0,0 @@
-#!/bin/bash
-set -e
-
-echo "==================================="
-echo "PaddleOCR-VL Server"
-echo "==================================="
-
-# Configuration
-MODEL_NAME="${MODEL_NAME:-PaddlePaddle/PaddleOCR-VL}"
-HOST="${HOST:-0.0.0.0}"
-PORT="${PORT:-8000}"
-MAX_BATCHED_TOKENS="${MAX_BATCHED_TOKENS:-16384}"
-GPU_MEMORY_UTILIZATION="${GPU_MEMORY_UTILIZATION:-0.9}"
-MAX_MODEL_LEN="${MAX_MODEL_LEN:-8192}"
-ENFORCE_EAGER="${ENFORCE_EAGER:-false}"
-
-echo "Model: ${MODEL_NAME}"
-echo "Host: ${HOST}"
-echo "Port: ${PORT}"
-echo "Max batched tokens: ${MAX_BATCHED_TOKENS}"
-echo "GPU memory utilization: ${GPU_MEMORY_UTILIZATION}"
-echo "Max model length: ${MAX_MODEL_LEN}"
-echo "Enforce eager: ${ENFORCE_EAGER}"
-echo ""
-
-# Check GPU availability
-if command -v nvidia-smi &> /dev/null; then
-    echo "GPU Information:"
-    nvidia-smi --query-gpu=name,memory.total,memory.free --format=csv
-    echo ""
-else
-    echo "WARNING: nvidia-smi not found. GPU may not be available."
-fi
-
-echo "Starting vLLM server..."
-echo "==================================="
-
-# Build vLLM command
-VLLM_ARGS=(
-    serve "${MODEL_NAME}"
-    --trust-remote-code
-    --host "${HOST}"
-    --port "${PORT}"
-    --max-num-batched-tokens "${MAX_BATCHED_TOKENS}"
-    --gpu-memory-utilization "${GPU_MEMORY_UTILIZATION}"
-    --max-model-len "${MAX_MODEL_LEN}"
-    --no-enable-prefix-caching
-    --mm-processor-cache-gb 0
-    --served-model-name "paddleocr-vl"
-    --limit-mm-per-prompt '{"image": 1}'
-)
-
-# Add enforce-eager if enabled (disables CUDA graphs, saves memory)
-if [ "${ENFORCE_EAGER}" = "true" ]; then
-    VLLM_ARGS+=(--enforce-eager)
-fi
-
-# Start vLLM server with PaddleOCR-VL
-exec vllm "${VLLM_ARGS[@]}"

test/helpers/docker.ts

@@ -49,7 +49,7 @@ export const IMAGES = {
 
   minicpm: {
     name: 'minicpm45v',
-    dockerfile: 'Dockerfile_minicpm45v',
+    dockerfile: 'Dockerfile_minicpm45v_gpu',
     buildContext: '.',
     containerName: 'minicpm-test',
     ports: ['11434:11434'],