fix image name

.gitea/workflows/build.yml

@@ -0,0 +1,53 @@
+name: Build and Push spark-vllm
+
+on:
+  push:
+    branches:
+      - main
+  workflow_dispatch:
+
+env:
+  IMAGE_NAME: spark-vllm-docker
+  IMAGE_TAG: latest
+
+jobs:
+  docker:
+    runs-on: dgx
+
+    steps:
+      - name: Checkout
+        uses: actions/checkout@v4
+
+      - name: Log in to Gitea Container Registry
+        uses: docker/login-action@v3
+        with:
+          registry: ${{ github.server_url }}
+          username: ${{ secrets.CI_USER }}
+          password: ${{ secrets.CI_PASSWORD }}
+
+
+      - name: Build image using upstream script
+        env:
+          DOCKER_DEFAULT_PLATFORM: linux/arm64
+        run: |
+          bash build-and-copy.sh -t ${IMAGE_NAME}:${IMAGE_TAG}
+
+      - name: Tag and Push Docker Image
+        shell: bash
+        run: |
+          VERSION=${{ github.run_number }}
+
+          REGISTRY=${GITHUB_SERVER_URL#https://}
+
+          TARGET_IMAGE=$REGISTRY/${{ github.repository_owner }}/${IMAGE_NAME}
+
+          docker tag \
+            ${IMAGE_NAME}:${IMAGE_TAG} \
+            $TARGET_IMAGE:$VERSION
+
+          docker tag \
+            ${IMAGE_NAME}:${IMAGE_TAG} \
+            $TARGET_IMAGE:latest
+
+          docker push $TARGET_IMAGE:$VERSION
+          docker push $TARGET_IMAGE:latest

README.md

@@ -1,4 +1,3 @@
-
 # vLLM Docker Optimized for DGX Spark (single or multi-node)
 
 This repository contains the Docker configuration and startup scripts to run a multi-node vLLM inference cluster using Ray. It supports InfiniBand/RDMA (NCCL) and custom environment configuration for high-performance setups.
@@ -1104,6 +1103,7 @@ The script attempts to automatically detect:
 *   **Cluster peers:** Discovered by scanning the `ETH_IF` subnet for hosts with SSH access **and** a GB10 GPU (`nvidia-smi --query-gpu=name` must return `NVIDIA GB10`).
 *   **Copy hosts (`COPY_HOSTS`):** In standard mode, same as cluster peers. In mesh mode, scanned separately on `enp1s0f0np0` and `enp1s0f1np1` subnets so that image/model transfers use the direct InfiniBand path.
 
+
 ### Manual Overrides
 
 You can override the auto-detected values if needed:

build-and-copy.sh

@@ -471,6 +471,7 @@ fi
 COMMON_BUILD_FLAGS+=("--build-arg" "BUILD_JOBS=$BUILD_JOBS")
 COMMON_BUILD_FLAGS+=("--build-arg" "TORCH_CUDA_ARCH_LIST=$GPU_ARCH_LIST")
 COMMON_BUILD_FLAGS+=("--build-arg" "FLASHINFER_CUDA_ARCH_LIST=$GPU_ARCH_LIST")
+COMMON_BUILD_FLAGS+=("--platform" "linux/arm64")
 if [ -n "$NETWORK_ARG" ]; then
     COMMON_BUILD_FLAGS+=("--network" "$NETWORK_ARG")
 fi