How to use FastFlowLM/Gemma3-4B-NPU2 with Transformers:
# Load model directly from transformers import AutoModel model = AutoModel.from_pretrained("FastFlowLM/Gemma3-4B-NPU2", device_map="auto")
