Any Doc to Explain How extract IQ1_S data from GGUFReader?

[ghostplant]

I want to read from gguf and do downstream inference, while I get these output fields from a IQS_1-quanted tensor:

ReaderTensor(name='blk.22.ffn_down_exps.weight', tensor_type=<GGMLQuantizationType.IQ1_S: 19>, shape=memmap([2048, 7168,  256], dtype=uint64), n_elements=3758096384, n_bytes=734003200, data_offset=48
603148224, data=memmap([[[ 94,  24, 191, ...,  63,  82,  74],
         [255,  21, 227, ..., 235, 230, 234],
         [185,  23, 138, ..., 209,  33, 115],
         ...,
         ...,
         [ 98,  24,  79, ..., 100, 143, 198],
         [141,  23, 244, ..., 253,  90,  94],
         [ 37,  24, 112, ..., 102, 106,  84]]],
       shape=(256, 7168, 400), dtype=uint8), field=ReaderField(offset=5260585, name='blk.22.ffn_down_exps.weight', parts=[memmap([27], dtype=uint64), memmap([ 98, 108, 107,  46,  50,  50,  46, 102, 102, 110,  95, 100, 111,
        119, 110,  95, 101, 120, 112, 115,  46, 119, 101, 105, 103, 104,
        116], dtype=uint8), memmap([3], dtype=uint32), memmap([2048, 7168,  256], dtype=uint64), memmap([19], dtype=uint32), memmap([48597887488], dtype=uint64)], data=[1, 3, 4, 5], types=[]))

Is there explanation of how each field is used in its dequantization?