Method quantize

utils/convert.py:84–85

(self, arr: NDArray)

    ggml_type: gguf.GGMLQuantizationType

    def quantize(self, arr: NDArray) -> NDArray:
        raise NotImplementedError(f'Quantization for {self.name} not implemented')

    def elements_to_bytes(self, n_elements: int) -> int:
        assert n_elements % self.block_size == 0, f'Invalid number of elements {n_elements} for {self.name} with block size {self.block_size}'