Skip to content

[atom-sgl-accuracy] Modify model cache mount #1352

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Open
junyyang-amd wants to merge 3 commits into
base: main
Choose a base branch
from
+12 −15
Open

[atom-sgl-accuracy] Modify model cache mount #1352

Changes from all commits
Commits
Show all changes
3 commits
Select commit Hold shift + click to select a range
874c982
Modify model cache mount
junyyang-amd Jun 25, 2026
f3ffe8a
Remove environment parameters about cache from container
junyyang-amd Jun 25, 2026
835e1b4
remove 'clean up containers'
junyyang-amd Jun 25, 2026
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
27 changes: 12 additions & 15 deletions .github/workflows/atom-sglang-accuracy-validation-gpu-shard.yaml
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -100,39 +100,39 @@ jobs:
- name: Prepare model cache mount
run: |
MODEL_CACHE_MOUNT=""
MODEL_CACHE_ROOT="/models"
MODEL_CACHE_ROOT=""
MODEL_CACHE_DESC=""

if [ -d "/shared/data/WRH/models" ]; then
MODEL_CACHE_ROOT="/shared/data/WRH/models"
MODEL_CACHE_MOUNT="-v ${MODEL_CACHE_ROOT}:${MODEL_CACHE_ROOT}"
MODEL_CACHE_MOUNT="-v ${MODEL_CACHE_ROOT}:/models"
MODEL_CACHE_DESC="${MODEL_CACHE_ROOT} (host mount)"
elif [ -d "/mnt/raid0/pretrained_model" ]; then
MODEL_CACHE_ROOT="/mnt/raid0/pretrained_model"
MODEL_CACHE_MOUNT="-v ${MODEL_CACHE_ROOT}:${MODEL_CACHE_ROOT}"
MODEL_CACHE_MOUNT="-v ${MODEL_CACHE_ROOT}:/models"
MODEL_CACHE_DESC="${MODEL_CACHE_ROOT} (host mount)"
elif [ -d "/data/pretrained_model" ]; then
MODEL_CACHE_ROOT="/data/pretrained_model"
MODEL_CACHE_MOUNT="-v ${MODEL_CACHE_ROOT}:${MODEL_CACHE_ROOT}"
MODEL_CACHE_MOUNT="-v ${MODEL_CACHE_ROOT}:/models"
MODEL_CACHE_DESC="${MODEL_CACHE_ROOT} (host mount)"
else
MODEL_CACHE_ROOT="/mnt/raid0/pretrained_model"
MODEL_CACHE_DESC="container-local ${MODEL_CACHE_ROOT} (no host cache mount)"
echo "Warning: Neither /mnt/raid0/pretrained_model nor /data/pretrained_model exists on runner; using container-local ${MODEL_CACHE_ROOT}."
echo "Warning: Neither /mnt/raid0/pretrained_model nor /data/pretrained_model nor /shared/data/WRH/models exists on runner; using container-local ${MODEL_CACHE_ROOT}."
fi

echo "Using model cache backend: ${MODEL_CACHE_DESC}"
echo "MODEL_CACHE_ROOT=${MODEL_CACHE_ROOT}" >> "$GITHUB_ENV"
echo "MODEL_CACHE_MOUNT=${MODEL_CACHE_MOUNT}" >> "$GITHUB_ENV"
echo "MODEL_CACHE_DESC=${MODEL_CACHE_DESC}" >> "$GITHUB_ENV"

- name: Clean up old containers
run: |
containers=$($CONTAINER_ENGINE ps -q)
if [ -n "$containers" ]; then
$CONTAINER_ENGINE kill $containers || true
fi
$CONTAINER_ENGINE rm -f "$CONTAINER_NAME" 2>/dev/null || true
#- name: Clean up old containers
# run: |
# containers=$($CONTAINER_ENGINE ps -q)
# if [ -n "$containers" ]; then
# $CONTAINER_ENGINE kill $containers || true
# fi
# $CONTAINER_ENGINE rm -f "$CONTAINER_NAME" 2>/dev/null || true

- name: Start validation container
run: |
Expand Down Expand Up @@ -166,9 +166,6 @@ jobs:
--ulimit stack=67108864 \
--env-file /tmp/sglang_env_file.txt \
-e HF_TOKEN="${HF_TOKEN:-}" \
-e HF_HOME="${MODEL_CACHE_ROOT}/.cache/huggingface" \
-e HUGGINGFACE_HUB_CACHE="${MODEL_CACHE_ROOT}/.cache/huggingface/hub" \
-e TRANSFORMERS_CACHE="${MODEL_CACHE_ROOT}/.cache/huggingface/transformers" \
--name "$CONTAINER_NAME" \
"${SGLANG_IMAGE_REF:-${SGLANG_IMAGE_TAG}}"
env:
Expand Down
Loading