Update Dockerfile

Dockerfile

@@ -18,8 +18,7 @@ USER user
 
 # Set environment variables
 ENV HOME=/home/user \
-    PATH=/home/user/.local/bin:$PATH \
-    PIP_USER=1
+    PATH=/home/user/.local/bin:$PATH
 
 # Set the working directory
 WORKDIR $HOME/app
@@ -27,11 +26,11 @@ WORKDIR $HOME/app
 # Copy the current directory contents into the container at /home/user/app
 COPY --chown=user . $HOME/app
 
-# Clone the aphrodite-engine from the specific branch
-RUN git clone --branch feat/exllamav2-support https://github.com/PygmalionAI/aphrodite-engine.git $HOME/aphrodite-engine
+# Install aphrodite-engine from PyPI to handle dependencies
+RUN python3 -m pip install aphrodite-engine
 
-# Install dependencies from the cloned repository
-RUN cd $HOME/aphrodite-engine && pip install -e .
+# Clone the specific branch of aphrodite-engine for the latest features
+RUN git clone --branch feat/exllamav2-support https://github.com/PygmalionAI/aphrodite-engine.git $HOME/aphrodite-engine
 
 # Install additional dependencies
 RUN pip install huggingface-hub hf-transfer
@@ -45,5 +44,5 @@ RUN huggingface-cli download LoneStriker/TinyLlama-1.1B-32k-Instruct-8.0bpw-h8-e
 # Expose the port the API server will listen on
 EXPOSE 7860
 
-# Command to run the API server
-CMD ["/bin/bash", "-c", "/bin/python3 -m aphrodite.endpoints.openai.api_server -q exl2 --dtype auto -gmu 0.95 --kv-cache-dtype fp8_e5m2 --max-num-seqs 15 --served-model-name "BagelMIsteryTour-v2-8x7B-AWQ" --enforce-eager -tp 4 --port 7860 --host 0.0.0.0 --model ~/goliath-gptq"]
+# Command to run the API server from the cloned directory
+CMD ["/bin/bash", "-c", "cd $HOME/aphrodite-engine && /bin/python3 -m aphrodite.endpoints.openai.api_server -q exl2 --dtype auto -gmu 0.95 --kv-cache-dtype fp8_e5m2 --max-num-seqs 15 --served-model-name "BagelMIsteryTour-v2-8x7B-AWQ" --enforce-eager -tp 4 --port 7860 --host 0.0.0.0 --model ~/goliath-gptq"]