How to use datasysdev/ann-sparseattention with Transformers:
# Load model directly from transformers import AutoModel model = AutoModel.from_pretrained("datasysdev/ann-sparseattention", dtype="auto")
