Skip to content

Latest commit

 

History

History
43 lines (33 loc) · 1004 Bytes

runbook.md

File metadata and controls

43 lines (33 loc) · 1004 Bytes

RUNBOOK

Run

Directly from image

$inpur_dir = "E:/ML-Data/video-transcription/input"
$output_dir = "E:/ML-Data/video-transcription/output"
$model_name = "base"
$device = "cuda"
docker run --gpus=all -i -t --env MODEL_NAME=${model_name} --env DEVICE=${device} -v ${inpur_dir}:/input -v ${output_dir}:/output hungdoan/video-transciption:latest

Using docker-compose

Using base model

docker compose -p "video-transcript" --env-file .env --env-file .env.base up --build --remove-orphans

Using largev3 model

docker compose -p "video-transcript" --env-file .env --env-file .env.largev3 up --build --remove-orphans

Build

Dev

docker buildx build --progress=plain --target dev .

Prod

Local build

docker buildx build --progress=plain --tag hungdoan/video-transciption:latest --target prod .

Push to Docker's registry

docker buildx build --target prod --tag hungdoan/video-transciption:latest --output=type=registry .