feat: re-enable EmbeddingGemma-300m support

Signed-off-by: Liav Weiss <[email protected]>

Author: Liav Weiss

.github/workflows/integration-test-docker.yml

@@ -82,18 +82,10 @@ jobs:
       - name: Download models
         run: |
           echo "Downloading minimal models for CI..."
-          # Authenticate with HuggingFace if token is available
-          # Note: For PRs from forks, HF_TOKEN is not available (GitHub security feature)
-          # The makefile will gracefully skip gated models (e.g., embeddinggemma-300m) if token is missing
-          if [ -n "${HF_TOKEN:-}" ]; then
-            huggingface-cli login --token "$HF_TOKEN" --add-to-git-credential
-            export HUGGINGFACE_HUB_TOKEN="$HF_TOKEN"
-          fi
           make download-models
         env:
           CI: true
           CI_MINIMAL_MODELS: true
-          HF_TOKEN: ${{ secrets.HF_TOKEN }}
           HF_HUB_ENABLE_HF_TRANSFER: 1
           HF_HUB_DISABLE_TELEMETRY: 1
           # HF_TOKEN is required for downloading gated models (e.g., embeddinggemma-300m)

tools/make/models.mk

@@ -169,7 +169,6 @@ download-models-full: ## Download all models used in local development and docs
 	@if [ ! -f "models/halugate-sentinel/.downloaded" ] || [ ! -d "models/halugate-sentinel" ]; then \
 		hf download LLM-Semantic-Router/halugate-sentinel --local-dir models/halugate-sentinel && printf '%s\n' "$$(date -u +%Y-%m-%dT%H:%M:%SZ)" > models/halugate-sentinel/.downloaded; \
 	fi
->>>>>>> fee395f (feat: re-enable EmbeddingGemma-300m support)
 
 # Download only LoRA and advanced embedding models (for CI after minimal tests)
 download-models-lora: