Add relocate, and a CUDA image for transcribing on an NVIDIA box
The database records absolute paths to the disc images, because audio, animation frames and transcription all read straight out of them. Carrying catalog.sqlite to another machine therefore leaves a catalogue that displays everything and can play nothing — and the same already applied between host and container, where the collection is /Users/... on one side and /media on the other. `relocate <directory>` re-points every copy, matching on filename and confirming by size, with --verify adding a full SHA-256 and --dry-run showing the outcome first. Copies that cannot be matched keep their old path and are reported rather than quietly rewritten. Names are compared after Unicode normalisation, which turned out to be the whole problem in practice: macOS stores "Wojna Trojańska.iso" decomposed (n + combining acute) while the database held it composed. Two of thirteen copies failed to match until that was fixed — and the same mismatch is exactly what would happen carrying a collection between macOS and Windows. Verified by rewriting the paths in a copy of the database to a Windows-shaped D:\Kolekcja, relocating, and then reading a voice line and an animation frame out of the discs through the relocated database. Dockerfile.cuda builds whisper.cpp with GGML_CUDA for compute capability 8.6 (GeForce RTX 30), on nvidia/cuda for both stages, with the JRE installed on top of the CUDA runtime. It is a separate file because both base images differ; folding it into the main Dockerfile would be more conditionals than content. Not verified beyond `docker build --check`: there is no NVIDIA GPU here, and the images are amd64. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
This commit is contained in:
Patryk Gensch
co-authored by
Claude Opus 5
Claude Opus 5
parent
b99f239553
commit
20ce26364c
@@ -0,0 +1,76 @@
|
||||
# Wariant pod NVIDIĘ — do transkrypcji na maszynie z kartą.
|
||||
#
|
||||
# Osobny plik, a nie przełącznik w Dockerfile, bo różnią się obrazy bazowe obu
|
||||
# etapów: whisper.cpp trzeba zbudować przy toolkicie CUDA, a obraz uruchomieniowy
|
||||
# musi nieść biblioteki CUDA i dopiero do nich dołożyć JRE. Wciśnięcie tego w
|
||||
# jeden plik dałoby więcej warunków niż treści.
|
||||
#
|
||||
# Budowanie i uruchomienie (wymaga NVIDIA Container Toolkit po stronie hosta;
|
||||
# na Windowsie: Docker Desktop z WSL2):
|
||||
# docker build -f Dockerfile.cuda -t rex-catalog:cuda .
|
||||
# docker run --rm --gpus all -v D:\Reksio:/media:ro -v katalog:/data \
|
||||
# -v D:\modele:/models:ro -e WHISPER_MODEL=/models/ggml-medium.bin \
|
||||
# rex-catalog:cuda transcribe
|
||||
#
|
||||
# Po przeniesieniu bazy z innej maszyny najpierw przypnij ją do kolekcji:
|
||||
# docker run --rm -v D:\Reksio:/media:ro -v katalog:/data rex-catalog:cuda relocate /media
|
||||
|
||||
FROM eclipse-temurin:21-jdk AS build
|
||||
WORKDIR /src
|
||||
|
||||
COPY gradlew settings.gradle build.gradle gradle.properties ./
|
||||
COPY gradle ./gradle
|
||||
COPY vendor ./vendor
|
||||
RUN ./gradlew --no-daemon dependencies --configuration runtimeClasspath > /dev/null 2>&1 || true
|
||||
|
||||
COPY src ./src
|
||||
RUN ./gradlew --no-daemon installDist \
|
||||
&& grep '^coreVersion=' gradle.properties | cut -d= -f2 > /src/core-version
|
||||
|
||||
# CMAKE_CUDA_ARCHITECTURES=86 to Ampere z GeForce RTX 30 (3060 ma 8.6).
|
||||
# Budowanie pod wszystkie architektury trwa wielokrotnie dłużej i puchnie,
|
||||
# a i tak nie przyda się na innej karcie bez przebudowy.
|
||||
FROM nvidia/cuda:12.6.2-devel-ubuntu24.04 AS whisper
|
||||
ARG CUDA_ARCH=86
|
||||
RUN apt-get update \
|
||||
&& apt-get install -y --no-install-recommends git cmake build-essential ca-certificates \
|
||||
&& git clone --depth 1 --branch v1.9.2 https://github.com/ggerganov/whisper.cpp /src \
|
||||
&& cmake -S /src -B /src/build -DCMAKE_BUILD_TYPE=Release -DBUILD_SHARED_LIBS=ON \
|
||||
-DGGML_NATIVE=OFF -DGGML_CUDA=ON -DCMAKE_CUDA_ARCHITECTURES=${CUDA_ARCH} \
|
||||
-DWHISPER_BUILD_TESTS=OFF -DWHISPER_BUILD_EXAMPLES=ON \
|
||||
&& cmake --build /src/build --target whisper-cli -j "$(nproc)" \
|
||||
&& mkdir -p /out/bin /out/lib \
|
||||
&& cp /src/build/bin/whisper-cli /out/bin/ \
|
||||
&& find /src/build \( -name 'libwhisper.so*' -o -name 'libggml*.so*' \) \
|
||||
-exec cp -P {} /out/lib/ \; \
|
||||
&& rm -rf /var/lib/apt/lists/* /src
|
||||
|
||||
FROM nvidia/cuda:12.6.2-runtime-ubuntu24.04
|
||||
WORKDIR /app
|
||||
|
||||
VOLUME ["/data"]
|
||||
ENV CATALOG_DATA=/data
|
||||
ENV WHISPER_BIN=/usr/local/bin/whisper-cli
|
||||
|
||||
# JRE z repozytoriów, bo obraz bazowy niesie CUDA, a nie Javę; ffmpeg do
|
||||
# przepróbkowania nagrań na 16 kHz mono, których whisper.cpp wymaga
|
||||
RUN apt-get update \
|
||||
&& apt-get install -y --no-install-recommends openjdk-21-jre-headless ffmpeg \
|
||||
&& rm -rf /var/lib/apt/lists/*
|
||||
|
||||
COPY --from=whisper /out/bin/whisper-cli /usr/local/bin/whisper-cli
|
||||
COPY --from=whisper /out/lib/ /usr/local/lib/
|
||||
RUN ldconfig
|
||||
|
||||
COPY --from=build /src/build/install/rex-catalog /app
|
||||
COPY --from=build /src/core-version /app/core-version
|
||||
COPY docker-entrypoint.sh /app/entrypoint.sh
|
||||
RUN chmod +x /app/entrypoint.sh \
|
||||
&& useradd --uid 10001 --create-home katalog \
|
||||
&& mkdir -p /data /media \
|
||||
&& chown katalog:katalog /data
|
||||
USER katalog
|
||||
|
||||
EXPOSE 8765
|
||||
ENTRYPOINT ["/app/entrypoint.sh"]
|
||||
CMD ["serve", "--host", "0.0.0.0"]
|
||||
Reference in New Issue
Block a user