lucasacchi/llamacpp-qwen3.5-0.8b
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

fix docker-compose and Dockerfile

Browse Source
This commit is contained in:
Luca Sacchi Ricciardi
2026-03-27 14:48:52 +00:00
parent f20f6571c2
commit 852481ea73
2 changed files with 3 additions and 4 deletions
Download Patch File Download Diff File Expand all files Collapse all files

5
Dockerfile
View File

@@ -5,20 +5,19 @@ ENV HSA_OVERRIDE_GFX_VERSION="11.0.0"
ENV AMDGPU_TARGETS="gfx1100" ENV AMDGPU_TARGETS="gfx1100"
ENV HIP_VISIBLE_DEVICES="0" ENV HIP_VISIBLE_DEVICES="0"
# Aggiunti hipblas-dev e rocblas-dev necessari per la compilazione GGML_HIP
RUN apt-get update && apt-get install -y \ RUN apt-get update && apt-get install -y \
build-essential cmake git libcurl4-openssl-dev \ build-essential cmake git libcurl4-openssl-dev \
hipblas-dev rocblas-dev \
&& rm -rf /var/lib/apt/lists/* && rm -rf /var/lib/apt/lists/*
WORKDIR /app WORKDIR /app
# Little often: cloniamo e compiliamo
RUN git clone https://github.com/ggerganov/llama.cpp.git . RUN git clone https://github.com/ggerganov/llama.cpp.git .
# Compilazione con supporto HIP nativo per AMD
RUN cmake -B build -DGGML_HIP=ON -DAMDGPU_TARGETS=gfx1100 -DCMAKE_BUILD_TYPE=Release RUN cmake -B build -DGGML_HIP=ON -DAMDGPU_TARGETS=gfx1100 -DCMAKE_BUILD_TYPE=Release
RUN cmake --build build --config Release -j $(nproc) RUN cmake --build build --config Release -j $(nproc)
EXPOSE 8000 EXPOSE 8000
# Punto d'ingresso nudo e crudo
ENTRYPOINT ["/app/build/bin/llama-server"] ENTRYPOINT ["/app/build/bin/llama-server"]

2
docker-compose.yml
View File

@@ -12,7 +12,7 @@ services:
# -ngl 99 offloada tutti i layer sulla GPU AMD # -ngl 99 offloada tutti i layer sulla GPU AMD
# --host 0.0.0.0 lo rende accessibile fuori dal container # --host 0.0.0.0 lo rende accessibile fuori dal container
command: > command: >
-m /models/qwen1_5-0_8b-chat-q8_0.gguf -m /models/Qwen3.5-0.8B-UD-Q8_K_XL.gguf
--host 0.0.0.0 --host 0.0.0.0
--port 8000 --port 8000
-ngl 99 -ngl 99