switch setup-10

Dockerfile

@@ -1,59 +1,44 @@
-# syntax=docker/dockerfile:1.3
-
 FROM ghcr.io/huggingface/text-generation-inference:3.3.4@sha256:91822fefc613d742f19f56b321925e81e41690dda29673c070fc68a65ae6bdb2
 
-# Install system dependencies and Python 3.11
+# Install system dependencies
 RUN apt-get update && \
     DEBIAN_FRONTEND=noninteractive apt-get install -y --no-install-recommends \
         build-essential git cmake ninja-build pkg-config curl ca-certificates gnupg software-properties-common \
         python3-pip python3.11 python3.11-dev python3.11-venv python3.11-distutils python3-dev tzdata && \
-    # Preconfigure timezone to UTC
     ln -fs /usr/share/zoneinfo/Etc/UTC /etc/localtime && dpkg-reconfigure --frontend noninteractive tzdata && \
-    # Add CUDA keyring and repository
     curl -fsSL https://developer.download.nvidia.com/compute/cuda/repos/ubuntu2204/x86_64/3bf863cc.pub | \
         gpg --dearmor --batch --yes -o /usr/share/keyrings/cuda-archive-keyring.gpg && \
     echo "deb [signed-by=/usr/share/keyrings/cuda-archive-keyring.gpg] https://developer.download.nvidia.com/compute/cuda/repos/ubuntu2204/x86_64 /" \
         > /etc/apt/sources.list.d/cuda.list && \
     apt-get update && \
     apt-get install -y --no-install-recommends cuda-compiler-11-8 && \
-    # Clean up
     rm -rf /var/lib/apt/lists/* && \
     ldconfig
 
-# Ensure python3 points to 3.11
+# Use Python 3.11 by default
 RUN update-alternatives --install /usr/bin/python3 python3 /usr/bin/python3.11 2
 
-# Upgrade pip and install PyTorch with CUDA support
+# Upgrade pip and install PyTorch
 RUN python3 -m pip install --upgrade pip && \
-    python3 -m pip install \
-        torch torchvision torchaudio --index-url https://download.pytorch.org/whl/cu118
-        
-# 1. Create a writeable cache directory
-RUN mkdir -p /tmp/hf_cache/hub
-
-# 2. Grant permissions so any user in the container can write there
-RUN chmod -R a+rwX /tmp/hf_cache
+    python3 -m pip install torch torchvision torchaudio --index-url https://download.pytorch.org/whl/cu118
 
-# 3. Point Hugging Face Hub to that cache
-ENV HF_HOME=/tmp/hf_cache \
-    HF_HUB_CACHE=/tmp/hf_cache/hub
-
-RUN \
-    # Clone diffvg recursively so we get the thrust submodule too
-    git clone --depth 1 --recursive https://github.com/BachiLi/diffvg.git /tmp/diffvg && \
+# Build and install diffvg (with thrust support)
+RUN git clone --depth 1 --recursive https://github.com/BachiLi/diffvg.git /tmp/diffvg && \
     cd /tmp/diffvg && \
-    # Replace bundled pybind11 with the released v2.12.1
+    # Replace internal pybind11, keep thrust submodule
     rm -rf pybind11 && \
     git clone --depth 1 --branch v2.12.1 https://github.com/pybind/pybind11.git pybind11 && \
     sed -i 's/^\s*cmake_minimum_required(VERSION .*$/cmake_minimum_required(VERSION 3.5)/' pybind11/CMakeLists.txt && \
-    # Configure & build without CUDA
     mkdir build && cd build && \
     cmake .. \
         -DCMAKE_BUILD_TYPE=Release \
         -DCMAKE_CXX_STANDARD=14 \
-        -DDIFFVG_CUDA=0 && \
+        -DDIFFVG_CUDA=0 \
+        -DTHRUST_INCLUDE_DIR=/tmp/diffvg/thrust && \
     make -j$(nproc) && \
-    # Install and clean up
     cd /tmp/diffvg && python3 setup.py install && \
     rm -rf /tmp/diffvg
 
+# Entrypoint remains unchanged
+ENTRYPOINT ["/usr/local/bin/text-generation-launcher"]
+