This repo contains a Dockerfile to run a peertube-runner instance. The generated image should be run as an Un-privileged user.
To build the image, run make build
. The build
task generates an image with the tag peertube-runner:latest
.
To build the image with the whisper-ctranslate2 for translation run
make build-whisper_ctranslate2
. This will build image with the tag
peertube-runner:latest-whisper_ctranslate2
.
The peertube-runner
instance can be configured using environment variable. The following environment variables
can be used :
PEERTUBE_RUNNER_JOBS_CONCURRENCY
: Amount of transcoding jobs to execute in parallel. Default:2
PEERTUBE_RUNNER_FFMPEF_THREADS
: Amount of threads used by ffmpeg for 1 local transcoding job. Default:2
PEERTUBE_RUNNER_FFMPEG_NICE
: ffmpeg niceness value, between -20 and 20. Default:0
PEERTUBE_RUNNER_REGISTERED_INSTANCE_URL
: A registered instance url. Default:none
PEERTUBE_RUNNER_REGISTERED_INSTANCE_RUNNER_TOKEN
: A registered instance runner token. Default:none
PEERTUBE_RUNNER_REGISTERED_INSTANCE_RUNNER_NAME
: A registered instance runner name. Default:none
PEERTUBE_RUNNER_TRANSCRIPTION_ENGINE
: The transcription engine to use. Default:none
PEERTUBE_RUNNER_TRANSCRIPTION_ENGINE_PATH
: The path to the transcription engine. Default:none
PEERTUBE_RUNNER_TRANSCRIPTION_MODEL
: The model to use for the transcription. Default:none
For the image build with whisper-ctranslate2 the default value change for:
PEERTUBE_RUNNER_TRANSCRIPTION_ENGINE
default towhisper-ctranslate2
PEERTUBE_RUNNER_TRANSCRIPTION_ENGINE_PATH
default to/usr/local/bin/whisper-ctranslate2
PEERTUBE_RUNNER_TRANSCRIPTION_MODEL
default totiny
$ docker run --rm -it \
-e PEERTUBE_RUNNER_REGISTERED_INSTANCE_URL=http://localhost:9000 \
-e PEERTUBE_RUNNER_REGISTERED_INSTANCE_RUNNER_TOKEN=foo-token \
-e PEERTUBE_RUNNER_REGISTERED_INSTANCE_RUNNER_NAME=instance-name \
peertube-runner:latest
The default command run in the container is peertube-runner server
which run a peertube-runner server.