search

meta-llama / llama-stack

Model components of the Llama Stack APIs
MIT License
3.68k stars 494 forks source link

missing target image architecture #253

Open nikolaydubina opened 3 hours ago

nikolaydubina commented 3 hours ago

right now image is build for same architecture as current host.

$ ./llama-stack/bin/llama stack build --template local --image-type docker --name llama-stack
$ docker image inspect llamastack-llama-stack | grep Architecture

        "Architecture": "arm64",

however, how can we specify different target architecture? if say we need linux/amd64?

picografix commented 2 hours ago

you can add platform here line:137 in llama_stack/distribution/build_container.sh

set -x
$DOCKER_BINARY build --platform "<platform>" $DOCKER_OPTS -t $image_name -f "$TEMP_DIR/Dockerfile" "$REPO_DIR" $mounts
nikolaydubina commented 1 hour ago

yes. that's what I ended up doing.

I am bringing this as GitHub Issue so hopefully this will be addressed in this tool.

picografix commented 1 hour ago

I have made some changes will raise a pr, just not sure where to add this change in documentation

On Wed, 16 Oct 2024 at 2:15 PM, Nikolay Dubina @.***> wrote:

yes. that's what I ended up doing.

I am bringing this as GitHub Issue so hopefully this will be addressed in this tool.

— Reply to this email directly, view it on GitHub https://github.com/meta-llama/llama-stack/issues/253#issuecomment-2416125196, or unsubscribe https://github.com/notifications/unsubscribe-auth/APBWYSKS25WBUZER4L4VOEDZ3YRUBAVCNFSM6AAAAABQAW3USCVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDIMJWGEZDKMJZGY . You are receiving this because you commented.Message ID: @.***>