# Native offline EmbeddingGemma 2 inference uses the verified GGUF worker.
# No sentence-transformers, PyTorch or model-download client is needed at runtime.
-r requirements.txt