v1.2.4

Signed-off-by: Marcel Klehr <mklehr@gmx.net>

Dockerfile

@@ -1,32 +1,32 @@
-FROM nvidia/cuda:12.2.2-cudnn8-runtime-ubuntu22.04
+FROM nvidia/cuda:12.4.1-runtime-ubuntu22.04
 
 ENV DEBIAN_FRONTEND=noninteractive
 
-RUN apt-get update && \
-    apt-get install -y \
-    software-properties-common
+RUN apt-get update
+RUN apt-get install -y software-properties-common
+RUN add-apt-repository -y ppa:deadsnakes/ppa
+RUN apt-get update
+RUN apt-get install -y --no-install-recommends python3.11 python3.11-venv python3-pip vim git pciutils
+RUN update-alternatives --install /usr/bin/python3 python3 /usr/bin/python3.11 1
+RUN apt-get -y clean
+RUN rm -rf /var/lib/apt/lists/*
 
-COPY requirements.txt /
+ENV NVIDIA_VISIBLE_DEVICES all
+ENV NVIDIA_DRIVER_CAPABILITIES compute
+ENV DEBIAN_FRONTEND=dialog
 
-ADD cs[s] /app/css
-ADD im[g] /app/img
-ADD j[s] /app/js
-ADD l10[n] /app/l10n
 ADD li[b] /app/lib
 ADD model[s] /app/models
 ADD default_confi[g] /app/default_config
 
-RUN add-apt-repository -y ppa:deadsnakes/ppa
-RUN apt-get update
-RUN apt-get install -y python3.11
-RUN apt-get install -y python3.11-venv
-RUN apt-get install -y python3.11-dev
-RUN apt-get install -y python3-pip
-
-ENV CMAKE_ARGS="-DLLAMA_CUDA=on"
-
-RUN \
-  python3 -m pip install -r requirements.txt && rm -rf ~/.cache && rm requirements.txt
+# Install requirements
+COPY requirements.txt /
+RUN python3.11 -m pip install --no-cache-dir --upgrade pip setuptools wheel
+RUN python3.11 -m pip install llama-cpp-python --extra-index-url https://abetlen.github.io/llama-cpp-python/whl/cu124
+# RUN python3 -m pip install -vvv --no-cache-dir https://github.com/abetlen/llama-cpp-python/releases/download/v0.2.81-cu124/llama_cpp_python-0.2.81-cp311-cp311-linux_x86_64.whl
+# https://github.com/abetlen/llama-cpp-python/releases/download/v0.2.81-cu122/llama_cpp_python-0.2.81-cp311-cp311-linux_x86_64.whl
+RUN sed -i '/llama_cpp_python/d' requirements.txt
+RUN python3.11 -m pip install --no-cache-dir --no-deps -r requirements.txt
 
 WORKDIR /app/lib
 ENTRYPOINT ["python3", "main.py"]

Makefile

@@ -21,7 +21,7 @@ help:
 .PHONY: build-push
 build-push:
 	docker login ghcr.io
-	docker buildx build --push --platform linux/amd64 --tag ghcr.io/nextcloud/llm2:1.1.0 .
+	docker build --push --platform linux/amd64 --tag ghcr.io/nextcloud/llm2:1.2.4 .
 
 .PHONY: download-models
 download-models:

appinfo/info.xml

@@ -9,7 +9,7 @@
 
 See [the nextcloud admin docs](https://docs.nextcloud.com/server/latest/admin_manual/ai/index.html) for more information.
 	]]></description>
-	<version>1.2.3</version>
+	<version>1.2.4</version>
 	<licence>MIT</licence>
 	<author mail="mklehr@gmx.net" homepage="https://github.com/marcelklehr">Marcel Klehr</author>
 	<author mail="bigcat88@icloud.com" homepage="https://github.com/bigcat88">Alexander Piskun</author>
@@ -25,7 +25,7 @@ See [the nextcloud admin docs](https://docs.nextcloud.com/server/latest/admin_ma
 		<docker-install>
 			<registry>ghcr.io</registry>
 			<image>nextcloud/llm2</image>
-			<image-tag>1.2.3</image-tag>
+			<image-tag>1.2.4</image-tag>
 		</docker-install>
 		<scopes>
 			<value>AI_PROVIDERS</value>