Reconfigured to use sh startup file.

Dockerfile

@@ -2,6 +2,7 @@ ARG UBUNTU_VERSION=22.04
 ARG CUDA_VERSION=12.3.1
 ARG BASE_CUDA_DEV_CONTAINER=nvidia/cuda:${CUDA_VERSION}-devel-ubuntu${UBUNTU_VERSION}
 ARG BASE_CUDA_RUN_CONTAINER=nvidia/cuda:${CUDA_VERSION}-runtime-ubuntu${UBUNTU_VERSION}
+ARG MODEL_URL=https://huggingface.co/IlyaGusev/saiga2_70b_gguf/resolve/main/ggml-model-q4_1.gguf
 
 FROM ${BASE_CUDA_DEV_CONTAINER} as build
 
@@ -22,11 +23,10 @@ ENV LLAMA_CUBLAS=1
 RUN mkdir build && \
     cd build && \
     cmake .. -DLLAMA_CUBLAS=ON && \
-#    cmake .. && \
     cmake --build . --config Release
 
 WORKDIR /data
-RUN wget https://huggingface.co/IlyaGusev/saiga2_70b_gguf/resolve/main/ggml-model-q4_1.gguf -nv -O model.gguf
+RUN wget ${MODEL_URL} -nv -O model.gguf
 
 FROM ${BASE_CUDA_RUN_CONTAINER} as runtime
 
@@ -36,7 +36,12 @@ WORKDIR /app
 # Copy the executable from the build stage
 COPY --from=build /build/llama.cpp/build/bin/server /app
 COPY --from=build /data/model.gguf /data/model.gguf
+COPY ./run.sh /app/run.sh
 WORKDIR /app
 EXPOSE 7860
 
-CMD ./server -m /data/model.gguf -c 4096 -ngl 70 --port 7860 --host 0.0.0.0
+# Make the script executable
+RUN chmod +x run.sh
+
+# CMD to run your script
+CMD ./run.sh

run.sh

@@ -0,0 +1,2 @@
+#!/bin/bash
+/app/server -m /data/model.gguf -c 4096 -ngl 70 --port 7860 --host 0.0.0.0